Investigation of Western Power’s low voltage …...Investigation of Western Power’s low voltage...
Transcript of Investigation of Western Power’s low voltage …...Investigation of Western Power’s low voltage...
Power Quality and Management Team
Investigation of Western Power’s low voltage operation limits with the integration of Photovoltaic systems
A report submitted to the School of Engineering and Energy, Murdoch University in partial fulfillment of the requirements for the degree of Bachelor of Engineering
By Kaveh Poyan
Student Number: 30639461
June 2010
i
Academic supervisor endorsement I am satisfied with the progress of this thesis project and that the attached report is an
accurate reflection of the work undertaken.
Dr Gregory Crebbin Date
ii
Abstract Distributed generation (DG), in the form of grid connected photovoltaic (PV) systems, is
expected to grow substantially throughout the South West Interconnected System (SWIS).
The growth of these systems is anticipated to meet a portion of the local energy
requirements and offset carbon emissions. The intermittency of the solar resource and its
relationship with conventional load and voltage management, presents challenges for
Western Power (the utility responsible for the SWIS). Therefore, investigating the effects
and management of grid connected PV systems in the SWIS, with respect to its voltage
operation limits, formed the primary objective of this project. The project studies were
performed on a typical Western Power (WP) low voltage (LV) network model, in
DIgSILENT Power Factory software, using WP residential network loadings data along
with solar radiation and temperature data. The power flow simulation results concluded that
PV penetration levels of up to 25% can be feasible in specific overhead (OH) networks.
Similarly, PV penetration levels of up to 49% can be sustained in particular underground
(UG) networks. Furthermore, these penetration levels were heavily dependant on existing
LV network balance and voltage profiles. The findings of the project also demonstrated that
grid connected PV systems are not offsetting the peaky SWIS residential network load
profile. Therefore, high grid connected PV system penetration, with no electrical storage
and demand management, provides substantially lower grid support value when compared
to dispatchable DG technologies.
iii
Acknowledgments I would like to thank my project supervisors, Dr Gregory Crebbin (Murdoch University)
and Nigel Wilmot (Western Power), for their enthusiasm and supervision of this thesis
project. Thanks also to Dr Trevor Pryor for his interests and input.
I would also like to acknowledge Western Power, and my team leaders (Abdul Haque and
Clayton Vander Schaaf) for their continuous support throughout my university studies.
iv
Table of contents
Academic supervisor endorsement.......................................................................................... i
Abstract .................................................................................................................................. ii
Acknowledgments.................................................................................................................iii
Table of contents ................................................................................................................... iv
Figures..................................................................................................................................vii
Tables .................................................................................................................................... ix
Acronyms ............................................................................................................................... x
1 Introduction .................................................................................................................... 1
1.1 Western Australia’s energy initiatives ................................................................... 1
1.2 Overview of distributed generation........................................................................ 2 1.2.1 Forms of DG....................................................................................................... 3
1.3 Changes to the Australian electrical utilities.......................................................... 4 1.3.1 Re-Structure and privatization ........................................................................... 4 1.3.2 Effect of deregulation on distributed generation................................................ 5 1.3.3 Environmental influence (on further change) .................................................... 5
1.4 Overview of PV systems ........................................................................................ 6 1.4.1 The off grid PV application................................................................................ 6 1.4.2 The grid connected PV systems ......................................................................... 6 1.4.3 PV market drivers............................................................................................... 8
1.4.3.1 Government policies .................................................................................. 9 1.4.3.2 Awareness of the renewable energy technologies...................................... 9
1.5 PV growth on the WP network ............................................................................ 10
1.6 Scope of project.................................................................................................... 11
1.7 Thesis outline ....................................................................................................... 13
2 The WA solar resource and its implications ................................................................ 14
2.1 Chapter overview ................................................................................................. 14
2.2 Solar characteristics.............................................................................................. 14 2.2.1 Review of the West Australian solar resource ................................................. 14
2.3 PV implications on the distribution grid .............................................................. 18 2.3.1 Potential benefits .............................................................................................. 18 2.3.2 Potential issues ................................................................................................. 19
2.3.2.1 Voltage operation limits ........................................................................... 19 2.3.2.2 Implications of LV network voltage variations for customers................. 21 2.3.2.3 Other potential issues ............................................................................... 22
3 Western Power’s distribution network......................................................................... 23
v
3.1 Chapter overview ................................................................................................. 23
3.2 Distribution HV parameters ................................................................................. 25
3.3 Distribution LV parameters.................................................................................. 27 3.3.1 Voltage operation limits ................................................................................... 27 3.3.2 The MEN system design .................................................................................. 27 3.3.3 LV network characteristics and sizes ............................................................... 28
3.3.3.1 Network size and feeder lengths .............................................................. 28 3.3.3.2 Network balance, zero sequence current.................................................. 29 3.3.3.3 District transformer tap settings ............................................................... 30 3.3.3.4 District transformer primary voltage........................................................ 31
4 Network modeling approach ........................................................................................ 32
4.1 DigSilent Power Factory software ....................................................................... 32 4.1.1 Distribution network component models ......................................................... 33
4.1.1.1 LV cable and conductor models............................................................... 33 4.1.1.2 Transformer model................................................................................... 33 4.1.1.3 Load models (single phase and three phase)............................................ 34 4.1.1.4 Inverter model .......................................................................................... 34 4.1.1.5 The primary HV network model .............................................................. 35
4.2 The specified network in Power Factory.............................................................. 36 4.2.1 Network selection criteria ................................................................................ 36 4.2.2 About the selected network .............................................................................. 36 4.2.3 After diversity maximum demand.................................................................... 39
4.3 LV load profile and scaling.................................................................................. 39 4.3.1 Load profile and scaling approach ................................................................... 41
4.4 PV array scaling factors ....................................................................................... 49 4.4.1 PV array scaling approach................................................................................ 50
5 Power flow simulation results and analysis ................................................................. 57
5.1 Chapter overview ................................................................................................. 57
5.2 Base case simulations and analysis ...................................................................... 57 5.2.1 Base case LV network with no PV penetration................................................ 57
5.2.1.1 Voltage profile results and discussions .................................................... 58 5.2.1.2 Grid losses ................................................................................................ 64
5.2.2 Base case LV network with PV penetration..................................................... 64 5.2.2.1 Voltage profile results and discussions .................................................... 65 5.2.2.2 Grid losses ................................................................................................ 69
5.3 CC4 sensitivity studies ......................................................................................... 70 5.3.1 Seasonal PV output and network load variations............................................. 72
5.3.1.1 Voltage profile results and discussions .................................................... 73 5.3.1.2 Network loadings and grid losses............................................................. 76
5.3.2 Feeder carrier type and network operation limits............................................. 78 5.3.2.1 7/4.75AAC Conductor voltage profile results and discussions................ 79 5.3.2.2 7/3.75AAC Conductor voltage profile results and discussions................ 81
vi
5.3.2.3 120SQmm XLPE LV cable voltage profile results and discussions........ 83 5.3.2.4 Grid losses comparison for the different carriers ..................................... 85
5.3.3 PV penetration and network balance variances................................................ 86 5.3.3.1 Three kW PV system investigations ........................................................ 86 5.3.3.2 PV systems on phase ‘a’ scaled to zero.................................................... 89 5.3.3.3 PV systems and load on phase b scaled to 1 and 0 .................................. 93
5.4 Summary of results............................................................................................... 98
6 Conclusion.................................................................................................................. 100
6.1 Conclusions ........................................................................................................ 100
6.2 Recommendations and future work.................................................................... 102 6.2.1 Single phase PV system connection for three phase customers..................... 102 6.2.2 LV network balance in new subdivisions ...................................................... 102 6.2.3 The effect of offloading or reverse power flow in district transformers........ 103 6.2.4 Demand management..................................................................................... 103
7 References .................................................................................................................. 104
Appendix A Load flow model ....................................................................................... 109
Appendix B Load and solar radiation data .................................................................... 110
Appendix C Power flow results..................................................................................... 111
vii
Figures Figure 1.1 Cumulative installed PV capacity between 1992 to 2008 in IEA PVPS
reporting countries [15].......................................................................................................... 7
Figure 1.2 Grid connected PV system growth on the SWIS [17].................................. 10
Figure 2.1 Average daily solar exposure across Australia [18]..................................... 15
Figure 2.2 Monthly average insolation incident on various surface tilts in WA
metropolitan [19].................................................................................................................. 16
Figure 2.3 Monthly averaged insolation normalized clearness index WA
metropolitan [19].................................................................................................................. 17
Figure 3.1 Western Power’s distribution network setup ............................................... 24
Figure 4.1 LV Cable and aerial conductor models ........................................................ 33
Figure 4.2 D/YN Transformer model ............................................................................ 34
Figure 4.3 Single phase Load model ............................................................................. 34
Figure 4.4 Three phase YN load model......................................................................... 34
Figure 4.5 External Grid model ..................................................................................... 35
Figure 4.6 The specified Low Voltage network in DIgSILENT Power Factory........... 37
Figure 4.7 Mean maximum temperature of Perth [42] .................................................. 40
Figure 4.8 Webb Street HV feeder lay out (dark purple) [17]....................................... 41
Figure 4.9 Monthly maximum, minimum and average Webb street feeder loadings ... 43
Figure 4.10 Scaled Webb St HV feeder loadings for July 2008...................................... 44
Figure 4.11 Average hourly temperature profile for July 2008....................................... 44
Figure 4.12 Scaled Webb St HV feeder loadings for August 2008................................. 45
Figure 4.13 Average hourly temperature profile for August 2008.................................. 45
Figure 4.14 Scaled Webb St HV feeder loadings for November 2008 ........................... 46
Figure 4.15 Average hourly temperature profile for November 2008............................. 46
Figure 4.16 Scaled Webb St HV feeder loadings for January 2009................................ 47
Figure 4.17 Average hourly temperature profile for January 2009 ................................. 47
Figure 4.18 Average hourly solar radiation for July 2008............................................... 51
Figure 4.19 Average hourly solar radiation for August 2008.......................................... 51
Figure 4.20 Average hourly solar radiation for November 2008 .................................... 52
viii
Figure 4.21 Average hourly solar radiation for January 2009......................................... 52
Figure 5.1 Base case Voltage profile CC1..................................................................... 59
Figure 5.2 Base case voltage profile CC2 ..................................................................... 59
Figure 5.3 Base case voltage profile CC3 ..................................................................... 60
Figure 5.4 Base case voltage profile CC4 ..................................................................... 60
Figure 5.5 Base case voltage profile CC5 ..................................................................... 61
Figure 5.6 Voltage profile CC1 with PV penetration .................................................... 65
Figure 5.7 Voltage profile CC2 with PV penetration .................................................... 66
Figure 5.8 Voltage profile CC3 with PV penetration .................................................... 66
Figure 5.9 Voltage profile CC4 with PV penetration .................................................... 67
Figure 5.10 Voltage profile CC5 with PV penetration .................................................... 67
Figure 5.11 CC4 voltage profile –July 2008, 1pm load and PV scales ........................... 73
Figure 5.12 CC4 voltage profile – August 2008, 1pm load and PV scales ..................... 74
Figure 5.13 CC4 voltage profile – November 2008, 1pm load and PV scales................ 74
Figure 5.14 CC4 voltage profile – January 2009, 1pm load and PV scales .................... 75
Figure 5.15 CC4 seasonal real and reactive power variations, with and without PV...... 76
Figure 5.16 CC4 seasonal variations in grid losses, with and without PV...................... 77
Figure 5.17 CC4 voltage profile, 7/4.75 conductor full load and no PV penetration...... 80
Figure 5.18 CC4 voltage profile, 7/4.75AAC low load and max PV output................... 80
Figure 5.19 CC4 voltage profile, 7/3.75AAC full load and no PV penetration .............. 82
Figure 5.20 CC4 voltage profile, 7/3.75AAC low load and max PV output................... 82
Figure 5.21 CC4 voltage profile, 120SQmm cable full load and no PV penetration ...... 84
Figure 5.22 CC4 voltage profile, 120SQmm cable, low load and max PV output ......... 84
Figure 5.23 Full load CC4 grid losses for the various carriers........................................ 86
Figure 5.24 CC4 Voltage profile, 3kW PV systems with 240SQmm cable.................... 87
Figure 5.25 CC4 Voltage profile, 3kW PV systems with 120SQmm cable.................... 87
Figure 5.26 CC4 Voltage profile, 3kW PV systems, no load with 120SQmm cable...... 88
Figure 5.27 CC4 Voltage profile, zero PV penetration on phase ‘a’, 120SQmm cable.. 91
Figure 5.28 CC4 Voltage profile, zero PV penetration on phase ‘a’, 7/4.75AAC .......... 91
Figure 5.29 CC4 Voltage profile, full PV penetration on phase ‘b’, 120SQmm cable ... 94
Figure 5.30 CC4 Voltage profile, full PV penetration on phase ‘b’, 7/4.75AAC ........... 94
ix
Tables Table 1.1 Cumulative installed PV capacity between 1992 to 2008 in IEA PVPS
reporting countries [15].......................................................................................................... 8
Table 2.1 Low voltage distribution system voltage limits [30]........................................ 21
Table 3.1 WP transformer list .......................................................................................... 28
Table 4.1 ADMD and land value [40].............................................................................. 39
Table 4.2 Selected LV network ADMD scaling factors .................................................. 48
Table 4.3 Selected solar radiation values and statistical analysis summary .................... 53
Table 4.4 Solar radiation and temperature data................................................................ 54
Table 4.5 Inverter output and scaling factors ................................................................... 55
Table 5.1 Base case district transformer and feeder parameter summaries ..................... 62
Table 5.2 Base case feeder zero sequence and neutral current summaries ...................... 62
Table 5.3 Base case grid losses ........................................................................................ 64
Table 5.4 CC4 phase loadings with PV............................................................................ 68
Table 5.5 Neutral current variations with and without PV............................................... 69
Table 5.6 Grid loss variances, with and without PV........................................................ 69
Table 5.7 CC4 installed load and PV system summary ................................................... 71
Table 5.8 Phase currents of CC4 at district transformer .................................................. 92
Table 5.9 Zero sequence currents of CC4 at district transformer .................................... 92
Table 5.10 Phase currents of CC4 at district transformer .............................................. 95
Table 5.11 Zero sequence currents of CC4 at district transformer ................................ 95
x
Acronyms SEI Strategic Energy Initiative
WA Western Australia
DG Distributed generation
PV Photovoltaic
WP Western Power
SWIS South West Interconnected System
AC Alternating Current
RAPS Hybrid remote area power supply
SHCP Solar Homes and Communities Plan
RET Renewable Energy Target
RECs Renewable Energy Certificates
LV Low Voltage
HV High voltage
OH Overhead
UG Underground
XLPE Cross linked polyethylene
RMS Root mean square
ADMD After Diversity Maximum Demand
AAC Aluminum Alloy Conductor
1
1 Introduction
1.1 Western Australia’s energy initiatives
The Strategic Energy Initiative (SEI) Issues Paper, released by the Government of Western
Australia (WA), Office of Energy, highlights the role of distributed generation (DG) in
meeting WA’s energy requirements in the next 20 years [1]. DG, in the form of grid
connected photovoltaic (PV) systems, has been identified to grow substantially throughout
the state, in meeting a portion of the local energy requirements and offsetting carbon
emissions [1]. In Western Power’s (WP) submission to the SEI issues paper, the need for
the growth of renewable DG forms has been acknowledged [2]. Moreover, WP recognizes
that the growth of such systems presents its benefits and challenges for the South West
Interconnected System (SWIS). The challenges in accommodating high DG penetration
levels in the SWIS can be attributed to the intermittency of renewable energy sources and
its relationship with load and voltage management [2].
WP is seeking to take a proactive approach in managing current and future DG penetration
levels in the SWIS, in line with the SEI goals over the next 20 years. As a result, this thesis
project has been initiated by the Power Quality Management Team of WP. Investigating
the effects and management of grid connected PV systems in the SWIS, with respect to its
voltage operation limits, forms the primary objective of this project.
2
1.2 Overview of distributed generation
Prior to progressing any further in this thesis project, it is important to establish common
understanding on the topic of DG. A literature review concluded that there is no fixed
definition to DG. However, some common characteristics can be listed as follows:
DG must be interconnected to the utility distribution grid [3-5]
DG is interconnected at or close to the load center [4, 6]
DG is small scale compared to centralized generation [4]
DG has gained momentum and renewed interest from industry over the past decade. This
can be attributed to technological innovations, the changing economics and the regulatory
environment. Some functions and advantages of DG are [3, 4, 7] :
If DG is used as backup generation, it can improve system reliability. The DG can
be brought on line when the grid supply is isolated or faulted.
DG can be connected in parallel with the utility service supply point. This may
provide benefits such as peak load reduction, voltage control and reactive power
compensation.
Integrating DG in a distribution electricity grid can present challenges for utility operators.
The issues are summarized [3, 4, 7]:
With the introduction of DG in a distribution system, two-way power flows will be
present and hence, power quality may be compromised as the existing distribution
network has been designed for radial power flow.
3
Safety could be compromised as the DG source may in some cases stay connected,
even when the main supply has failed (islanding). This could mean back feed in to a
faulted or isolated distribution line.
Network short circuit characteristics change with the integration of DG and this can
cause protection systems to lose coordination.
1.2.1 Forms of DG
Many sources [3, 8] point out that distributed generation is not necessarily renewable
generation. Some common forms of DG are [3, 6, 8, 9]:
Reciprocating Engines
Gas turbines
Micro Turbines
Fuel cells
Photovoltaic
Wind
Other renewables – Thermal solar, small hydro, geothermal, wave, etc
Given the nature of renewable sources such as wind and solar, it is most feasible to use
these as small scale DG sources rather than large centralised plants.
4
1.3 Changes to the Australian electrical utilities
1.3.1 Re-Structure and privatization
Deregulation and restructuring of the electricity industry is a major change that has
occurred in the last decade, not only in Australia but around the world as well. Prior to
1994, nearly the entire Australian electricity market consisted of vertically integrated state
monopolies [10]. In the early 1990s the state owned Victorian electricity market was
privatized. The state of South Australia was second to privatize its electricity market
during the years of 1999 – 2000 [10]. During this time all the other jurisdictions have also
disaggregated their formerly integrated industry, although Tasmania and Western Australia
have so far retained government ownership.
The drivers for the deregulation of the Australian electricity market, according to literature
can be attributed to [10, 11]:
The recognition that other countries were achieving considerably greater
efficiencies than Australia in electricity supply;
National Competition Policy (NCP) involving a general review of the operations of
“essential facilities” (which were, in the main, owned by governments) and a
requirement that they be opened to non-affiliates on reasonable terms; and
Consequences of poor financial circumstances in the States of Victoria and South
Australia resulting in new governments which sold its energy assets partly in pursuit
of a privatization agenda and in part to reduce debt.
5
In the disaggregated Australian electricity market, independent generators inject power into
the grid. Retailers purchase wholesale electricity from the suppliers and sell to the
consumers who are free to choose their energy supplier.
1.3.2 Effect of deregulation on distributed generation
Advancements in Alternating Current (AC) grid technology have led to large scale
generation, transmission and distribution grids. Conventional generation included thermal
plants and nuclear plants. However, there is consensus [3, 8, 9] that of late an increasing
fraction of generation is embedded within distribution systems. The restructuring of the
electricity industry is leading to an increased interest in distributed generation. This is
because DG can potentially delay expensive transmission and distribution reinforcement
projects [12].
1.3.3 Environmental influence (on further change)
The predominant fuel used for the production of electricity in Australia is coal [10]. The
consequences of using coal is large amounts of pollutants being released into the
atmosphere, which under the future regulatory environment and carbon emission targets
would not be permitted. Hence, there is a need to use an increasing mix of renewable
energy to meet energy requirements. This has been recognized by the WP’s Submission to
the SEI issues paper [2].
6
1.4 Overview of PV systems
1.4.1 The off grid PV application
PV systems are used for off grid and on grid applications. Although the off grid application
does not concern this project, some general information with regards to these systems has
been provided.
The off grid PV application is generally used to meet energy requirements in conjunction
with battery banks. These are used in developed and developing countries. Some of the
applications for this configuration include [13, 14]:
Lighting
Pumping water
Communications (telephone, facsimile, radio) and other electrical devices
Small devices (calculators)
Remote community power supply
Hybrid remote area power supply (RAPS) systems
1.4.2 The grid connected PV systems
Most grid connected PV systems are connected in parallel with the distribution grid’s
supply point to meet electrical power requirements of a building. This implies that when
the installed PV array is generating power and in the same building one or more devices are
simultaneously consuming power, the devices are at least partially powered by the PV
system without using the utility grid supply. The surplus of solar energy is fed into the
utility distribution grid and bought by the respective energy retailer.
7
Grid connected PV systems have grown substantially over the past decade world wide, as
can be seen in figure 1.1. This trend is evident in the Australian context as displayed in
Table 1.1.
Figure 1.1 Cumulative installed PV capacity between 1992 to 2008 in IEA PVPS reporting countries [15]
8
Table 1.1 Cumulative installed PV capacity between 1992 to 2008 in IEA PVPS reporting countries [15]
1.4.3 PV market drivers
The rapid growth of PV systems in Australia can be attributed to the following key market
drivers [13]:
Government policies
Price of electricity
Awareness of the technology and RE desire
These are discussed in more detail in the following sections.
9
1.4.3.1 Government policies
Government rebates for residential PV systems under the Solar Homes and Communities
Plan (SHCP) prior to middle of 2007 were AUD4000 for the first kW installed. This value
was increased to AUD8000 which ignited 4.6MW of PV to be installed in 2007. The SHCP
PV rebate has been replaced with Solar Credits Scheme as of 9th June 2009 as part of the
Renewable Energy Target (RET) scheme of 450,000 GWh by 2020. Solar Credits will be
provided in the form of Renewable Energy Certificates (RECs) for new solar PV systems
installed. The Solar Credits will apply to the first 1.5 kW of capacity of the system
installed. The level of support provided under this scheme will depend on the market value
of RECs, subject to variation over time, and the location and size of the installed system.
For example, based on a $30 REC price in late 2009, a solar PV system in Newcastle,
Sydney, Perth, Adelaide, Brisbane or Canberra will receive [16]:
1.0 kW Solar PV System: $3,090 (103 RECs)
1.5 kW Solar PV System: $4,650 (155 RECs)
2.0 kW Solar PV System: $4,950 (165 RECs)
2.5 kW Solar PV System: $5,250 (175 RECs)
3.0 kW Solar PV System: $5,550 (185 RECs)
1.4.3.2 Awareness of the renewable energy technologies
The awareness of climate change through main stream media as well as improved
government rebates for PV installations and variable grid feed in tariffs has increased
public interest in the solar energy technology and installations. PV technology is also used
by industry and government groups to develop an environmentally friendly, sustainable and
10
responsible image [13]. This trend of PV growth is expected to continue as Solar City
installations across Australia begin, as building energy standards improve and as energy
prices increase due to international resource prices, infrastructure upgrades and emissions
trading [2].
1.5 PV growth on the WP network
The number of customers applying to connect PV systems to the SWIS has increased in the
past few years due to the aforementioned drivers. This can be seen in figures 1.2, where
each brown triangle represents a request to connect a small scale renewable system, mainly
PV, to the WP distribution grid.
Figure 1.2 Grid connected PV system growth on the SWIS [17]
11
The WP distribution grid is based on radial power flow design, hence, voltage is highest at
the distribution transformer and lowest at the end of the distribution feeder(s). The
combination of radial power flow design, low load factors and the volatile PV output,
means there is a need for specific research in the area of voltage operation limits under high
DG penetration scenarios.
As the number of PV systems connecting to the SWIS is expected to continue to grow, the
effects of these systems on the WP distribution grid need to be studied in order to ensure
safety, reliability and power quality of electricity is not compromised. WP needs to
establish technical rules and follow Australian Standards in order to accommodate grid
connected PV systems on the distribution grid. It is also necessary to realize any benefits of
such systems through field experience, simulations and research.
1.6 Scope of project
The objectives of this project are outlined:
1. Investigate the WP Low Voltage (LV) network voltage operation limits with the
introduction of PV systems, by considering the variations in PV system
performances and network loadings conditions.
2. Provide background on the WP distribution network design philosophy with respect
to their voltage operation limits and its control.
3. Create a comprehensive distribution network model in DIgSILENT Power Factory
(power flow simulation software) in line with the WP design philosophies.
12
4. Identify feasible PV penetration levels on the WP LV networks, with respect to
voltage operation limits and perform sensitivity analyses.
The reduction of LV network losses with the introduction of PV systems has also been
considered in this project. In addition, based on the findings of specific network setup
scenarios, insight into issues surrounding three phase customers with single phase PV
system installations are discussed succinctly.
A WP LV residential network model consisting of many single phase customers was
created in DIgSILENT Power Factory. Residential distribution network loading data was
used in determining LV network loadings for various times, in the period of June 2008 to
May 2009. Similarly, solar radiation data was obtained, adjusted and used in conjunction
with daily temperature data to estimate PV system performances throughout that year. The
network loading and PV performance estimates were embedded in the DIgSILENT Power
Factory model and used for the investigations.
The assessment of viability in terms of economics of PV systems is not considered in this
project. As stated earlier, this project concentrates on investigating critical PV penetration
levels with respect to the WP distribution LV network voltage profiles.
13
1.7 Thesis outline
This thesis consists of seven chapters and electronic appendices. Chapter one gives insight
into the growth of DG in the form of grid connected PV systems on the SWIS. It also
outlines the objectives of this project. In chapter two, an assessment of the WA solar
resource is presented. The potential benefits and implications of PV systems on the WP
distribution network are discussed in detail.
Chapter three aims to describe the factors that affect the WP high voltage (HV) and LV
distribution voltage operation limits. The allowable voltage operation limits and the design
considerations of the HV and LV distribution network are stated.
Chapter four steps through the approach taken to model the WP distribution network based
on the WP design philosophy. Details of the methods used to derive the network load
model and PV systems outputs have also been outlined in this chapter.
Chapter five presents the results of the power flow simulations carried out on the specified
network model. There, the sensitivity study results on various network parameters are
presented. The chapter ends with a summary of the network study findings.
Chapter six outlines the important findings of the project. It also provides some scope and
questions that need to be addressed by WP in the future.
14
2 The WA solar resource and its implications
2.1 Chapter overview
In this chapter the WA solar resource is assessed and its characteristics are defined.
Furthermore, the potential benefits and implications of PV systems on the WP distribution
network are discussed.
2.2 Solar characteristics
Understanding the West Australian solar characteristics is an important step in identifying
the effects of PV penetration on the WP distributing network. The solar characteristics of a
region determine the expected energy yield of such systems, including the seasonal
variations in this yield. This, combined with the existing distribution network
characteristics, would determine the overall interaction of specific networks with PV
systems.
2.2.1 Review of the West Australian solar resource
Figure 2.1 displays the annual average daily global solar exposure over Australia for the
period 1990 to 2008. The annual average daily global solar radiation for WA is between
5.0-5.8 kWh per square metre.
15
Figure 2.1 Average daily solar exposure across Australia [18]
In figure 2.2, it can be seen that the monthly averaged solar radiation values vary for
different surface tilts, with the slope of 32 and17 degrees resulting in the maximum annual
energy yield of 5.8 kWh/m2. Figure 2.2 shows that, by adjusting the tilt of the north facing
surface, the seasonal variations in the output of a PV array can be adjusted. This is a result
of the apparent position of the sun in the WA skies over a year, where higher slopes have
higher output in the winter months and lower slopes have higher output in summer.
16
Monthly Averaged Insolation incident on various surfaces
0
1
2
3
4
5
6
7
8
9
Ann
ual
Ave
rage
Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Annual
Month of the year
Sola
r ins
olat
ion
(kW
h/m
.squ
are/
day
Tilt 0Tilt 32Tilt 47Tilt 17
Figure 2.2 Monthly average insolation incident on various surface tilts in WA metropolitan [19]
Figure 2.3 displays the normalized clearness index values. Clearness Index is the ratio of
horizontal solar radiation at the earth’s surface to extraterrestrial solar radiation on a
horizontal surface. The values in the chart range from 0.48 for the winter months and 0.64
for the summer months. The clearness index values displayed in figure 2.3 indicate a high
percentage of clear sky days in the summer months and moderately clear sky days in the
winter months.
17
Monthly Averaged Insolation Normalized Clearness Index (0 to 1.0)
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec
Month of the year
Cle
arne
ss in
dex
Clearness index
Figure 2.3 Monthly averaged insolation normalized clearness index WA
metropolitan [19]
The annual average solar radiation values shown in figures 2.1 and 2.2, along with the
clearness index displayed in figure 2.3, indicate that WA has an excellent solar resource.
Although there is some seasonal variation in solar radiation, it can easily be
counterbalanced by varying the tilt of the PV array. In summary, the WA solar resource has
the following distinct characteristics:
Maximum solar radiation occurs in the months of November to February.
Minimum solar radiation occurs in the months of May to August.
Optimum annual solar radiation is achieved from tilted surfaces with values close to
the latitude angle.
Based on the aforementioned points, it is apparent that the maximum PV penetration on the
distribution network will occur at around solar noon, throughout the year. Studies to be
18
conducted in the later sections of this project will specifically be constructed around these
distinct characteristics.
2.3 PV implications on the distribution grid
2.3.1 Potential benefits
Grid connected PV systems can have many positive effects on an electrical utility grid.
Many sources share this view and some common points of interest with these systems
include [2, 20-25]:
Improved system voltage profiles
Reduced grid power losses
Deferred upgrades for an existing infrastructure (feeders, transformers and
switchgear)
Decreased transmission and distribution related costs
Improved reliability
Reduced carbon emission levels
Adopting the technology means higher understanding of power consumption is
developed and hence greater attention to a rational use of the energy in general and
to environmental problems.
In order to achieve the aforementioned benefits, PV penetration levels and connection
points may need to be controlled [23]. Although the grid connected PV penetration levels
can be controlled via network assessments and an approval process, the location and
orientation of such systems cannot be optimized to the needs of a distribution network. This
19
is because these are mainly residential customer driven projects. However, in WA,
generally the benefits may be enhanced and well pronounced due to the great solar resource
of this region.
2.3.2 Potential issues
2.3.2.1 Voltage operation limits
The introduction of PV systems to a radial designed distribution network can compromise
voltage operation limits, often in the form of network voltage rise [26-29]. Voltage
operation limits in the context of this project corresponds to voltages outside the nominated
allowable limits, under or over. Traditionally, distribution networks in Australia and
worldwide have been designed to have a maximum voltage set point at the transformer
node. This voltage set point made allowances for a radial voltage drop due to the series
impedance of the network toward the end of the feeder, ensuring voltage limits are
satisfied. The existing distribution network is now facing the possibility of two way power
flows across feeders and even district transformers, depending on PV and other forms of
DG penetration levels. This means that at times, the voltage set point at the district
transformer is required to vary in order for voltage operation limits to be met.
Voltage operation limits in LV distribution networks, with the introduction of single and
three phase grid connected PV systems, can be affected due to factors including:
The off loading of feeders as a consequence of clustered PV system penetration
results in a reduced voltage drop across the series impedance of LV networks and
hence a local voltage rise [29].
20
Uneven distribution of single phase PV systems in a three phase LV network will
result in higher neutral currents and phase voltage imbalances.
Seasonal and daily solar radiation variations result in a variable PV output. This
means changes to the distribution network voltage profile throughout the year.
Challenges are presented here as LV distribution network voltage control methods
are passive.
The maximum solar radiation and hence PV system output may not be matched
with the network’s ‘peaky load period’ but rather coincide with minimum loadings
[29]. This would enhance the network voltage variations [2].
Ambient temperature affects a normal residential load profile as well as varying the
PV system output, hence variations in network voltage profile will be observed.
The points outlined above illustrate the combined effects of the dynamic nature of load
profiles, existing distribution network shape and solar radiation output on network voltage
operation limits. For instance, a warm and comfortable sunny day will result in minimum
loading of LV feeders combined with maximum PV penetration (due to improved PV
performance at lower temperature). This in turn results in a smaller voltage drop across the
series impedance of a LV feeder and hence, a voltage rise. Indeed, this voltage rise will
vary at different points of the network, depending on factors such as existing cable or
overhead (OH) line parameters, combined with network load balance across the three
phases.
21
The challenge here is to accommodate the maximum number of PV systems on the passive
distribution network and maintain network voltages within limits. This means that peak PV
output and its daily and seasonal variations need to be determined and modeled amongst
typical load profiles and its variation. These network characteristics will be discussed,
specifically in the context of the WP distribution network design, in the next chapter of this
thesis project.
2.3.2.2 Implications of LV network voltage variations for customers
Network Voltage operation limits can have significant effects on WP customers. Apart
from the risk of appliance failures due to over or under voltages, customers who have PV
systems connected to the LV distribution network may experience disconnection time from
the distribution grid. WP requires the grid connected inverters to stay connected to the
network for voltage ranges listed in table 2.1, where the specified voltage ranges are 5-
minute averages of the root mean square (RMS) value.
Table 2.1 Low voltage distribution system voltage limits [30]
This table implies that where LV network voltages are outside the specified limits, be it as
a result of the grid connected inverters or not, the inverters will trip, resulting in lost
22
revenue for the customers. This issue of lost revenue has been further aggravated by the
fact that the Australian Standard, AS 4777, requires grid connected inverters to disconnect
from the distribution network within 2 seconds, should the network voltage rise above
270V and below 200 volts [31]. The other issue is that the inverters do not measure average
5-minute RMS values but rather instantaneous or 5-second average values, meaning more
frequent tripping of the inverter under the WP allowable voltage limits. In brief, it is
essential that the network voltage is not adversely affected as a result of PV system
connections.
2.3.2.3 Other potential issues
Other potential issues associated with the introduction PV systems to a distribution grid,
but not concerning this project include [21, 23, 32]:
Islanding (where PV systems back feed power in to an isolated or faulted line)
Total Harmonic Distortion of voltage and current waveforms
Protection and coordination between network devices
The anti-islanding schemes that are deployed in the Australian approved inverters
(according to AS 4777 [31]) are accurate and responsive. On the whole, the factors
mentioned above are secondary issues that have been addressed via research, state of the art
power electronic technologies and relevant national and international standards.
23
3 Western Power’s distribution network
3.1 Chapter overview
This chapter aims to describe the factors that affect HV and LV distribution voltage
operation limits. Understanding Western Power’s distribution network design philosophy is
a prudent step in deriving a representative network model in power flow simulation
software, such as Power Factory. The key design aspects of the WP network, concerning
this project are:
Distribution high voltage configuration is delta.
Distribution Low voltage configuration is wye.
LV distribution network makes use of Multiple Earthed Neutral (MEN) earthing
strategy.
LV network voltages at district substations are preset and can only be adjusted
manually after transformer de-energisation.
Various cables and conductors are used throughout the distribution network
Figure 3.1 shows an overview of the WP distribution network set up.
24
Figure 3.1 Western Power’s distribution network setup
25
3.2 Distribution HV parameters
Distribution HV networks distribute power from zone substations to district substations.
These networks are designed to safely distribute apparent power in the order of 3MVA – 15
MVA [33] while keeping losses at a minimum. The WP HV distribution network consists
of feeders rated at 6.6kV, 11kV, 22kV and 33kV of delta configuration. Metropolitan
distribution networks distribute power to district substations at 6.6kV, 11kV and 22kV
while rural networks make use of 22kV and 33kV feeders. 22kV feeders are most common
for both rural and metropolitan distribution networks.
The HV distribution network is required to operate within voltage limits of +/- 10% of the
nominal network voltage [30]. The HV network voltages are kept within these design
voltage limits by making use of different types of conductors and cables at various points
of the network. The parameters of the carriers used, along with the nature of the feeder
loads, determine the feeder voltage profile at various points of the network. This voltage
profile in essence is a function of:
Feeder thermal limits (the series impedance of the feeder)
Distance from the zone substation (feeder length)
Feeder load balance
The complex line current value (hence real and reactive power delivered)
It is common that the voltage set point for these feeders is highest at the zone substations
and lowest towards the end of the feeders.
26
Where applicable, WP makes use of voltage regulators and reactive power compensation to
control and keep the feeder voltages within the voltage design criteria. Feeders that make
use of voltage control and compensation devices are generally rural feeders, which have the
following characteristics:
Are of long lengths (extend furthest from zone substations)
Are well loaded (to voltage capacity limits of the feeder)
Consist of unbalanced phase loads with single phase spurs
Metropolitan HV distribution feeders generally have good load balance across the three
phases, are of shorter lengths and hence have superior voltage profiles compared to rural
feeders. As a result, these feeders are constrained due to their thermal limits rather than
voltage capacity and design limits.
With the increasing PV penetration across the SWIS, it is likely that the HV network
voltage profiles will also be affected, and indeed, the changes will vary for different feeders
according to their characteristics. For the purpose of this project, the HV voltage set points
and variations have not been taken in to account specifically but rather in the form of the
district transformer tap settings and its input voltage.
27
3.3 Distribution LV parameters
3.3.1 Voltage operation limits
WP’s LV network distributes power from district substations to customers at a nominated
line to line voltage of 415V with wye configuration. The steady state voltage operation
limits are defined [30]:
± 6% of the nominal voltage during normal operating state,
± 8% of the nominal voltage during maintenance conditions,
±10% of the nominal voltage during emergency conditions.
These voltage operation limits are currently controlled via radial LV network design, tap
changing district transformers (manual operation) and the use of various cables and
conductors.
3.3.2 The MEN system design
In a MEN system, the neutral conductor of the distribution system provides the low
impedance return path for zero sequence currents due to each phase. The potential above
earth of this neutral conductor is kept low by a sufficient number of earth connections
throughout its length. As the neutral conductor is connected to exposed metal frames of
appliances etc, it is prudent that the potential difference between earth and the neutral wire
is kept to a minimum [34]. In brief, the neutral conductor is an important aspect of the WP
distribution system that needs appropriate attention in establishing representative network
models.
28
3.3.3 LV network characteristics and sizes
Western Power operates many district substations in order to ensure LV customers are
serviced safely and reliably. The following sections seek to explain some key WP network
characteristics with respect to network and feeder voltage profiles.
3.3.3.1 Network size and feeder lengths
The LV distribution network size is very much dependant on the size of the respective
district transformer. Table 3.1 lists the different transformers that exist within the WP
distribution network.
Table 3.1 WP transformer list
TX size kVA Phase Technology O / U Zone 10 1 OH R 25 1 OR 3 OH R / SR 50 1 UG R / SR 63 3 UG / OH R / SR 100 3 OH R / SR / M 200 3 OH R / SR / M *300 3 UG / OH M 315 3 UG M *500 3 OH M 630 3 OH M
1000 3 OH M
R= rural, SR=semi rural and M=metropolitan.
OH=overhead and UG=underground.
* indicates transformer is no longer available for network design
29
The common transformer sizes within the metropolitan areas, for the residential
underground (UG) networks, are 500kVA and 630kVA. The 1000kVA transformers are
predominantly used in commercial networks. WP makes use of 120mm2, 185mm2 and
240mm2 Cross Linked Polyethylene insulated (XLPE) cables to supply customers from the
district substations. The voltage profile of the LV networks is dependant on the total load,
the length and the cables that make up a given feeder.
The 100kVA and 200kVA OH transformers usually are configured as OH meshed
networks in metropolitan, semi rural and rural areas, with the main variance between these
networks being the load disparity. For example, the loads within a metropolitan LV
network will be compact in comparison with the rural networks due to the density of the
dwellings. The OH meshed networks makeup most of the WP aged infrastructure. Due to
the natural load growth, unbalanced loading of the phases and the occasional existence of
poor conductors, these networks have weak voltage profiles. The feeder lengths in these
networks can be limited due to their voltage profile dropping below the allowable
operational limits as well as the size of the OH transformers.
It is therefore prudent to model different load balances, cable and conductor sizes in
simulations of LV networks under PV penetration scenarios.
3.3.3.2 Network balance, zero sequence current
The WP LV distribution network consists of many single phase customers. These single
phase loads cause imbalances within the three phase LV network and hence, the result is a
30
zero sequence current flow due to each phase in the neutral wire. This neutral current then
results in a potential difference across the series impedance of the neutral carrier which
varies at different points of the LV network [34]. In the case of large phase loading
imbalances of a LV network, the neutral current may get quiet high. This would produce
voltage drops across long neutral wires as well as serious line to neutral network voltage
imbalances. These voltage imbalances can be in the form of a phase to neutral voltage rise
in one phase and a phase to neutral voltage dip in another.
Based on the above analysis, it is necessary to model the neutral wire of the WP LV system
appropriately. It is as important to model single phase loads connected to phase and neutral
as the impedance of the neutral carrier and its voltage drop are of significance.
3.3.3.3 District transformer tap settings
The district transformers used in the WP distribution network inherit five tap settings on
their primary windings. Tap three is used where the nominal transformer primary voltage is
present and the secondary nominal voltage of 440V, 1.06pu is desired. With the nominal
district transformer primary voltage, taps one and two step the secondary voltage down in
steps of 2.5% (maximum of 5%) while taps 4 and 5 step the secondary voltage up in steps
of 2.5% (maximum of 5%).
In illustration of the use of these tap settings, a long LV feeder that is well loaded should be
considered. It is expected that the voltages towards to the end of the feeder will be close to
the lower allowable limits. Therefore, the district substation transformer secondary voltage
31
can be set close to the upper allowable operation limit (440V or 1.06pu) in order to cater
for the series volt drop across the feeders or vice versa.
It is therefore important to model a default LV network transformer tap setting that enables
all feeder loads to be supplied while voltages remain within operation limits. It is also vital
to monitor the change in LV network voltage profile with maximum possible PV
penetration and minimum network loads.
3.3.3.4 District transformer primary voltage
As discussed in section 3.2 of this report, the HV distribution feeder voltage varies at
different points along a feeder. This in turn means that the primary voltage of the district
transformers at various points of the distribution networks also is subject to variations,
depending on its location along the HV feeder route. These variations are once again dealt
with by making use of the district transformer tap settings. This implies the transformers
closest to the zone substations need to adopt higher primary to secondary turn ratios (hence
taps one and two) compared to those situated further from zone substations. Thus, it is
important to make an assumption on the primary voltage of a district substation, and to
keep this constant throughout all simulations in order to monitor LV distribution network
voltage profiles with the introduction of PV systems.
32
4 Network modeling approach This chapter describes the approach taken to model the WP distribution network based on
the WP design philosophy, which was discussed in the chapter 3 of this report. DIgSILENT
Power Factory version 14 was used for this task, and hence, an overview of this software is
presented in this chapter. The network component models created within DIgSILENT
Power Factory to represent the WP distribution network, will also be discussed briefly.
Details of the methods used to derive the network load models and PV systems outputs will
be outlined as well.
4.1 DigSilent Power Factory software
The program Power Factory has been written by DIgSILENT and is a computer aided
engineering tool that is used for the analysis of industrial, utility and commercial electrical
power systems [35]. The DIgSILENT Power Factory program was selected to model the
distribution network for the purpose of this project as it is the standard program used in WP
and Murdoch University for conducting detailed power flow studies. The software is
capable of modeling basic and more complex power systems, depending on the user
requirements and knowledge of power systems.
For the purpose of attaining meaningful results, this project has aimed to create a
comprehensive WP distribution network model in DIgSILENT Power Factory software. In
the endeavor of achieving this task, specific network components based on the WP network
33
design philosophy as discussed, in chapter 3 of this report, have been created in the
program. These are discussed in detail in the next section of this report.
4.1.1 Distribution network component models
The following network component models have been created in Power Factory. It should
be noted that these models are consistent with the WP network diagram displayed in figure
3.1 of this report.
4.1.1.1 LV cable and conductor models
Some of the various cable and conductor models that are used in the WP distribution
network have been created in Power Factory. These have a continuous neutral conductor as
displayed in figure 4.1. Please refer to appendix A for carrier details.
Figure 4.1 LV Cable and aerial conductor models
4.1.1.2 Transformer model
The 500kVA transformer model is shown in figure 4.2. The transformer is of delta / wye
neutral configuration and has its star point connected to ground and the neutral wire as per
the MEN system design. The tap settings of the transformer have also been implemented in
this model, where tap 3 is the nominal position. The higher primary to secondary turns ratio
taps are represented by taps 1 and 2 and the lower primary to secondary turns ratios
represented by taps 4 and 5. Please refer to appendix A for transformer details.
34
Figure 4.2 D/YN Transformer model
4.1.1.3 Load models (single phase and three phase)
The single phase and three phase load models are displayed in figures 4.3 and 4.4
respectively. The single phase loads are connected across phase to neutral, as per the WP
LV network arrangement.
Figure 4.3 Single phase Load model
Figure 4.4 Three phase YN load model
4.1.1.4 Inverter model
Grid connected inverters inject power in to LV networks at unity power factor [36]. Hence,
in steady state analysis it is acceptable to model the grid connected inverters as PQ
elements, that is, as a negative load with injected current [23, 37]. The inverter model is
35
therefore based on that of the single phase load model, injecting power at unity power
factor, as displayed in figure 4.3.
4.1.1.5 The primary HV network model
The HV network side of the transformer is required to have a line to line voltage of 22kV,
with delta configuration. As a result, Power Factory’s external grid element was used to
model the transformer’s primary network. This is shown in figure 4.5. The voltage set point
here can be adjusted according to the requirements of the user. In the case of this project,
the voltage set point remains fixed at 1pu.
Figure 4.5 External Grid model
36
4.2 The specified network in Power Factory
4.2.1 Network selection criteria
In order to carry out the required power flow simulations to investigate the LV distribution
voltage operation limits under PV penetration, a LV distribution network needed to be
created in Power Factory. Sensitivity studies on the LV network voltage operation limits
could then be carried out under worst case scenarios by varying:
Main feeder carriers
Amount of PV penetration on the feeder
Network balance
By performing the above sensitivity studies, insight into the voltage profile of different
network sizes can be gained. Based on this idea and the aim to keep this project to a
manageable size, it was decided to model one operational WP LV network in Power
Factory. Furthermore, although the base case power flow and voltage profiles of the
specified network would be investigated, it was decided to conduct the sensitivity studies
only on one of the long and well loaded LV feeders.
4.2.2 About the selected network
The selected WP LV distribution network is shown in figure 4.6. This network consists of a
well loaded 500kVA district transformer and five LV feeders supplying predominantly
single phase residential customers [38]. The primary voltage for the selected network is
22kV. Other network data, such as LV feeder lengths and load technologies, have been
extracted from the original network design [38].
37
CC4
C4-J19-L1 or A1
Circuit #Junction #
Or Load #
PV array #
Note: PV symbol longer than load symbol
Load
Figure 4.6 The specified Low Voltage network in DIgSILENT Power Factory
38
The existing LV feeders denoted CC (circuit) in figure 4.6 are all 240SQmm XLPE cables
of various lengths. The customer services are mainly supplied via junctions with 25SQmm
XLPE cable tee offs from the 240SQmm XLPE cables. Other customer service connections
include those which have direct connections with the 240SQmm XPLE feeders. The Power
Factory model shown in figure 4.6 is in appendix A of this report.
Information on the network phase which the single phase customers are connected to is not
electronically available. However, it is a WP requirement to evenly distribute single phase
loads across the three phases of a network [39]. Therefore, the single phase loads in the
selected LV network have been distributed as evenly as possible across the three phase
feeders.
The After Diversity Maximum Demand (ADMD) for the domestic loads of the LV network
is 4.7kVA per lot [40]. The power factor of the loads has been assumed to be 0.9 and
lagging.
All the introduced PV systems to the LV network are single phase and have a rating of
1.5kW (according to the new Solar Credit Scheme promoting this size). They are all
connected to the customer’s main switch board, and hence, export excess power in to the
respective LV network phase, through the service carrier. A brief analysis of 3kW PV
systems at every customer connection point will also be made. Detailed investigation of the
1.5kW PV system spread in LV networks is sufficient at this stage, as it is unlikely for all
residential customers on a feeder to have systems larger than 1.5kW installed in near future.
39
4.2.3 After diversity maximum demand
The maximum demand on a transformer or a LV feeder, when divided by the number of
loads supplied, provides a value which is in essence the “average contribution per
customer”, or simply the “average demand” for a typical customer [39]. This value varies
between 4.7kVA to 8.7kVA depending on the suburb and hence price of land, as displayed
in table 4.1.
Table 4.1 ADMD and land value [40]
Lot Price ($) Single to Quradruplex $512,000 or less 4.7kVA 512 to 1,024,000 6.2kVA Above 1,024,000 8.7kVA
Given that the average price of land in WA is less than $512,000 [41], the specified
network ADMD value of 4.7kVA is a good representation of the residential services across
the SWIS. As a result, the specified ADMD values are assumed to be the ultimate
maximum load for the services specified in the LV network under investigation, and that
they can be scaled down according to seasonal variations.
4.3 LV load profile and scaling
Residential LV network loadings are always subject to variations. In general these
variations are a function of time of the day and season of the year. Feeder loading varies
with time of day, due to the consumer’s life style, and time of year, due to cooling and
heating requirements. In illustration of this point, the mean maximum temperature of Perth
Metropolitan areas has been displayed in figure 4.7.
40
Figure 4.7 Mean maximum temperature of Perth [42]
Here, it can be seen that the months of March to May and October to November produce
moderate temperatures and hence, house hold heating and cooling requirements are at a
minimum. This implies that consumers are not coming home and switching on air
conditioners or heaters, and in comparison different load profiles are expected for the
summer and winter months.
Based on the above discussion, an approximate approach needs to be followed in order to
establish typical LV domestic feeder loadings at different times of the day and year. This
approach has been based on the scaling of typical residential ADMD values according to a
normalized representative load profile of a residential HV distribution feeder.
41
4.3.1 Load profile and scaling approach
The Webb Street HV feeder, which sources from Riverton Zone Substation, consists of
93% residential customers [17]. As a result, it was decided to attain scaling factors for the
selected LV network loads from this HV feeder, which has been displayed in figure 4.8.
Figure 4.8 Webb Street HV feeder lay out (dark purple) [17]
The Webb Street HV feeder loading data for the period of June 2008 to May 2009 was
obtained from WP [43]. These feeder loadings were average half hour current readings.
The daily hourly maximum, average and minimum feeder loadings for each month were
plotted using Microsoft excel. This was done to allow for good visibility of feeder loadings
42
over the seven days of the week for each month. Please refer to appendix B for loading
profiles and raw data.
After analyzing the Webb Street feeder loading plots in detail, it was decided to extract
scaling factors for the specified LV network from the Webb Street feeder loadings for the
months of July 2008, August 2008, November 2008 and January 2009. The month of July
was selected as it produced a typical winter load profile and hence the winter LV network
voltage operation limits could be studied. August was interesting in the sense that it
produced high solar radiation readings during 2008, see the next section. November
loadings were low: with good solar radiation (see figure 4.20) and relatively low
temperature values, as shown in figure 4.7, it was clear that the PV systems would perform
well. Therefore, it was deemed necessary to analyse the LV network voltage operation
limits for this month. January was chosen as it produced a typical summer load profile and
high solar radiation readings.
The absolute maximum, minimum and average feeder loading values for each month are
displayed in figure 4.9, where it can be seen that a maximum reading of 285A was recorded
for March 2009.
43
Max, Min and average feeder loadings
0.00
50.00
100.00
150.00
200.00
250.00
300.00
Apr-08 Jun-08 Jul-08 Sep-08 Oct-08 Dec-08 Feb-09 Mar-09 May-09 Jul-09
Month
Feed
er c
urre
nt (A
)
Max (A)Min (A)Ave (A)
Figure 4.9 Monthly maximum, minimum and average Webb street feeder loadings
The half hour feeder loadings for the selected months were then normalized according to
the absolute maximum. Plots of the normalized daily average hourly feeder loadings for the
selected months are displayed in figures 4.10, 4.12, 4.14 and 4.16. Instantaneous ten minute
air temperature readings were also downloaded from Murdoch University’s On Line
Weather Station website [44]. These have been plotted as daily average hourly temperature
values and can be used to observe the effect of temperature on feeder loadings. The daily
average hourly temperature plots are shown in figures 4.11, 4.13, 4.15 and 4.17.
The legend in the plots presented in the following figures, display Monday as 1 to Sunday
as 7.
44
Scaled Webb St HV feeder loadings for July 2008
0
0.1
0.2
0.3
0.4
0.5
0.6
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
Hour of day
Feed
er n
orm
alis
ed c
urre
nt v
alue
1234567
Average of Normalised Jul08
hour of day
Day Jul08
Figure 4.10 Scaled Webb St HV feeder loadings for July 2008
Temperature profile for July 2008
0.000
2.000
4.000
6.000
8.000
10.000
12.000
14.000
16.000
18.000
20.000
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
Hour of day
Tem
pera
ture
(C d
eg)
1234567
Average of T Jul-08
Hour of day
Day Jul-08
Figure 4.11 Average hourly temperature profile for July 2008
45
Scaled Webb St HV feeder loadings for August 2008
0
0.1
0.2
0.3
0.4
0.5
0.6
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
Hour of day
Feed
er n
orm
alis
ed c
urre
nt v
alue
1234567
Average of Normalised Aug08
hour of day
Day Aug08
Figure 4.12 Scaled Webb St HV feeder loadings for August 2008
Temperature profile for August 2008
0.000
5.000
10.000
15.000
20.000
25.000
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
Hour of day
Tem
pera
ture
(C d
eg)
1234567
Average of T Aug-08
Hour of day
Day Aug-08
Figure 4.13 Average hourly temperature profile for August 2008
46
Scaled Webb St HV feeder loadings for November 2008
0
0.05
0.1
0.15
0.2
0.25
0.3
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
Hour of day
Feed
er n
orm
alis
ed c
urre
nt v
alue
1234567(blank)
Average of Normalised nov08
hour of day
Day Nov08
Figure 4.14 Scaled Webb St HV feeder loadings for November 2008
Temperature profile for November 2008
0.000
5.000
10.000
15.000
20.000
25.000
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
Hour of day
Tem
pera
ture
(C d
eg) 1
234567(blank)
Average of T Nov-08
Hour of day
Day Nov-08
Figure 4.15 Average hourly temperature profile for November 2008
47
Scaled Webb St HV feeder loadings for January 2009
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
Hour of day
Feed
er n
orm
alis
ed c
urre
nt v
alue
1234567
Average of Normalised Jan09
hour of day
Day Jam09
Figure 4.16 Scaled Webb St HV feeder loadings for January 2009
Temperature profile for January 2009
0.000
5.000
10.000
15.000
20.000
25.000
30.000
35.000
40.000
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
Hour of day
Tem
pera
ture
(C d
eg)
1234567
Average of T Jan-09
Hour of day
Day Jan-09
Figure 4.17 Average hourly temperature profile for January 2009
48
The effect of temperature on residential network loadings becomes apparent when
assessing the feeder loading data and temperature profiles displayed in the figures above.
The winter load profile produces distinct morning and evening peaks. The load profile of
November displays low network loadings throughout the day due to the moderate
temperatures. January is producing high network loadings in the afternoons and evenings,
due to the high temperatures.
Some carefully selected mean average hourly normalized HV feeder loadings from the
selected months were then used as hourly scaling factors for the specified LV network
loads. These scaling factors were selected such that they coincided with the maximum solar
radiation readings. The extracted scaling factors from each month are listed in table 4.2.
Statistical analysis was carried out on these scaling factors to show variations and standard
error, in the mean hourly values of the selected months. For full details, see appendix B.
Table 4.2 Selected LV network ADMD scaling factors
Month Jul Aug Nov Jan Hour of day 12 13 13 16 10 13 17 9 13 17 Mean feeder scaling factor 0.27 0.27 0.27 0.27 0.22 0.24 0.27 0.29 0.39 0.51Minimum 0.24 0.2 0.24 0.25 0.21 0.23 0.25 0.26 0.31 0.45Maximum 0.33 0.3 0.32 0.3 0.23 0.25 0.28 0.33 0.46 0.58Standard deviation 0.03 0 0.02 0.02 0.01 0.01 0.01 0.03 0.06 0.05
These scaling factors have been entered as “load characteristics” in Power Factory in order
to scale all the single phase ADMDs in one easy step according to the power flow
requirements. Some important points to note, with regards to the extraction and use of these
scaling factors for power flow simulations, are outlined below:
49
These LV network scaling factors provide possible and realistic network loadings
for different days of the year and times of the day.
They provide hourly average LV network loadings which are respectively lower
and higher than the ultimate maximum and minimum feeder loadings.
It is not a requirement to analyse the LV network power flow studies for all the
scaling factors, but rather for those that give the worst case feeder voltage profile
scenarios. (Hence, low loads and max PV system performance)
One limitation of these scaling factors is that they do not scale the feeder loads to
zero or full load. Another limitation is that they are scaling the entire feeder loads at
the same time.
In order to overcome the limitations of these scaling factors, power flow studies with the
network load scaled to one and zero may also be conducted, as these are other realistic LV
network loading conditions.
4.4 PV array scaling factors
PV systems have power output ratings in accordance with Standard Test Conditions (STC).
STC involve a solar radiation level of 1000Wmsq and a module temperature of 25°C [45].
If solar radiation levels are lower than 1000Wmsq and or the module temperature is above
25°C, the output of a PV module, and hence the array, will be reduced. Even though the
installed PV arrays have nominal ratings, the actual performance of these systems is very
much dependant of the amount of solar radiation incident on the array as well as the
ambient temperature of the modules.
50
For PV installations in the field, here in WA, STC are unlikely to occur and as discussed in
section 2 of this report solar radiation is subject to seasonal and daily variations. Further, in
assessing figure 4.17, it is clear that temperatures in WA are likely to be a lot higher than
that of STC during the summer months, when solar radiation is also high. Therefore, for the
purpose of this project, it is required to attain the necessary data and to take an approximate
approach in predicting the output of PV arrays.
4.4.1 PV array scaling approach
In order to estimate the likely seasonal and daily variations about the nominal rating of PV
arrays in WA, ten minute average horizontal short wave radiation flux density (J m-2 s-1)
data for the period of June 2008 to May 2009 were downloaded from Murdoch University’s
On Line Weather Station website [44]. As per the Webb Street feeder data, the daily hourly
maximum and average solar radiation readings for each month were plotted using
Microsoft excel. Please refer to appendix B for solar radiation plots and raw data.
It was necessary to monitor and extract solar radiation data from the months of July,
August, November and December, in line with the discussions in the previous section of
this report. The daily average hourly solar radiation readings for these months are displayed
in figures 4.18-4.21.
The legend in the plots presented in the following figures, display Monday as 1 to Sunday
as 7.
51
Solar radiation profile for July 2008
0.000
50.000
100.000
150.000
200.000
250.000
300.000
350.000
400.000
450.000
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
Hour of day
Sola
r rad
iatio
n (W
)
1234567
Average of SR Jul-08
Hour of day
Day Jul-08
Figure 4.18 Average hourly solar radiation for July 2008
Solar radiation profile for August 2008
0.000
100.000
200.000
300.000
400.000
500.000
600.000
700.000
800.000
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
Hour of day
Sola
r rad
iatio
n (W
)
1234567
Average of SR Aug-08
Hour of day
Day Aug-08
Figure 4.19 Average hourly solar radiation for August 2008
52
Solar radiation profile for November 2008
0.000
200.000
400.000
600.000
800.000
1000.000
1200.000
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
Hour of day
Sola
r rad
iatio
n (W
) 1234567(blank)
Average of SR Nov-08
Hour of day
Day Nov-08
Figure 4.20 Average hourly solar radiation for November 2008
Solar radiation profile for January 2009
0.000
200.000
400.000
600.000
800.000
1000.000
1200.000
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
Hour of day
Sola
r rad
iatio
n (W
)
1234567
Average of SR Jan-09
Hour of day
Day Jan-09
Figure 4.21 Average hourly solar radiation for January 2009
53
The selected average hourly solar radiation data from each month are listed in table 4.3.
Statistical analysis was carried out on these values to show variations and standard error for
selected data. For full details, see appendix B.
Table 4.3 Selected solar radiation values and statistical analysis summary
Month Jul Aug Nov Jan Hour of day 12 13 13 16 10 13 17 9 13 17 Mean horizontal SR (W/msq) 301 300 626 400 727 863 374 564 939 535 Minimum SR (W/msq) 192 215 552 335 594 770 310 472 860 490.6Maximum SR (W/msq) 383 417 703 457 840 954 459 640 1049 572.7Standard deviation 61 71 57 47 103 69 49 62 64 33
It is understood that PV array installations across WA have a slope close to the latitude
angle as this results in the highest annual energy yield from these systems. Therefore, the
extracted horizontal average hourly solar radiation data listed in table 4.3 had to be
converted to values for a slope of 32 degrees. The method described by Christiana
Honsberg & Stuart Bowden [45] was used to convert the horizontal solar radiation readings
to that of a north facing surface with a slope of 32 degrees. Comparison between the
calculated mean solar radiation values and long term average hourly solar radiation data,
for a surface tilt of 32 degrees was made. The long term data was collected from the
Australian Solar Radiation Data Handbook [46]. The results are shown in table 4.4. The
mean average hourly temperature data have been extracted from figures 4.11, 4.13, 4.15
and 4.17.
54
Table 4.4 Solar radiation and temperature data
Month Jul Aug Nov Jan Hour of day 12 13 13 16 10 13 17 9 13 17 Mean 2008/2009 32deg (W/msq) 469 468 868 555 703 834 374 535 891 507 Long term mean 32deg (W/msq) 631 611 697 371 757 935 293 539 999 403 Difference (W/msq) -162 -143 171 184 -54 -101 81 -4 -108 104 Mean temperature (°C) 16 16 18 18 20 22 21 24 30 28
The mean hourly solar radiation data for the period of 2008/2009 and the respective mean
temperature values in table 4.4, have been used to predict the output of north facing PV
arrays, with a tilt angle equal to the latitude, under the following assumptions:
The inverter efficiency is 87% for medium and low loading periods.
The inverter efficiency is 92% for higher loading periods.
The module temperature remains at 20°C above mean ambient conditions.
The effect of temperature on the power output of PV modules is 0.5 %/°C.
Equations 1 and 2 were then used in predicting the output of the specified PV systems.
(( ) ) ( ) ⎟⎠⎞
⎜⎝⎛×⎥⎦
⎤⎢⎣⎡ ×−+−×=
100010025201 SRDF
TAA ModaNomout (1)
Invoutout AInv η×= (2)
Where,
Aout = PV array output in Watts
ANom = PV array nominal rating
Ta = Ambient Temperature
55
DFMod = Derating factor for the effect of temperature on the power output of PV modules in
%
SR = Solar radiation incident on the PV array
Invout = Output from the inverter in Watts
The estimated inverter outputs, for a north facing 1.5kW array have been listed in table 4.5.
Table 4.5 Inverter output and scaling factors
Month Jul Aug Nov Jan Hour of day 12 13 13 16 10 13 17 9 13 17 Inverter efficiency (%) 87 87 92 87 87 92 87 87 92 87 Inverter output (W) 578 577 1120 677 849 1053 449 632 1076 586 Array scaling factor 0.39 0.38 0.75 0.45 0.57 0.70 0.30 0.42 0.72 0.39
The specified PV array outputs have been scaled with respect to the 1.5kW array size and
entered in Power Factory, as “PV array scaling factors”. It should be noted that these
scaling factors will correctly scale the output of various PV system sizes. Some important
points to note in regards to the use of these scaling factors include:
They provide realistic PV array output values through grid connected inverters.
However, the exact output of a system at any point in time is subject to the
respective atmospheric conditions.
It is not a requirement to analyse the LV network power flow studies for all the
scaling factors but rather for those thought to give the worst case feeder voltage
profile scenarios.
As displayed in table 4.4, long term data shows that the ultimate maximum solar
radiation for arrays tilted at 32 degrees may reach value higher than those for the
56
specified time period of June 2008 to May 2009. Hence, power injection to the LV
distribution grid from grid connected PV systems may exceed those resulting from
the highest scaling factors listed in table 4.4.
In order to estimate the possible maximum PV system power injection to the LV
distribution grid, power flow studies with the arrays scaled to 80% of nominal need to be
conducted.
57
5 Power flow simulation results and analysis
5.1 Chapter overview
In this chapter, the results and analysis of the load flow studies, conducted in DIgSILENT
Power Factory, for the various network conditions are presented. As specified in section
4.2.1 of this report, the power flow studies conducted for the entire specified LV network
(figure 4.6) formed the base case analysis with the network having no PV penetration. Base
case power flow studies have been conducted in order to establish the existing network
operation limits and parameters. These could then be used as a means of monitoring
changes to network operation limits and parameters as a result of PV penetration.
Power flow studies and network analysis has been carried out for all the specified network
loads having a 1.5kW grid connected PV system. Thereafter, the sensitivity study results
on various parameters, for the long and highly loaded feeder denoted CC4 (circuit 4) is
presented. The chapter ends with a concise summary of the network study findings.
5.2 Base case simulations and analysis
5.2.1 Base case LV network with no PV penetration
In section 3.3.3 of this report, it was recognized that the district transformer’s primary
voltage and tap settings directly affect the LV network voltage profile and operation limits.
As per those discussions, the district transformer’s primary voltage for the specified
network was decided to be set to 1.0pu (22kV), with the tap setting of 3, which in turn
resulted in a no load secondary voltage of 1.06pu. These base case network operating
58
configurations were kept consistent through out the rest of the power flow studies, and
where required variations to these settings have been recommended.
In order to establish the existing network voltage operation limits, power flow studies with
the network load and PV output characteristics set to one and zero respectively were
conducted. This was done to make certain all the LV feeder voltages (CC1 to CC5) were
within operation limits under the full load conditions. Other network parameter results,
such as LV feeder currents, power flows, power factor, grid losses and descriptive voltage
profiles have been documented and can be used for analysis at various points of the
network. For the full results of this power flow study scenario, please refer to appendix C
of this report.
5.2.1.1 Voltage profile results and discussions
The voltage profiles for the feeders in the specified LV network, namely CC1 to CC5 are
represented in figures 5.1-5.5. The voltage profile plots display the line to ground voltage
magnitudes as a function of distance away from the low voltage frame, which is directly
connected to the secondary of the district transformer.
59
Figure 5.1 Base case Voltage profile CC1
Figure 5.2 Base case voltage profile CC2
60
Figure 5.3 Base case voltage profile CC3
Figure 5.4 Base case voltage profile CC4
61
Figure 5.5 Base case voltage profile CC5 Looking at figures 5.1 to 5.5, it is clear that the feeder voltages are within the specified
voltage operation limits, stated in section 3.3.2 of this report. It should also be noted that
the voltage set point at the low voltage frame is not 1.06u. This can be attributed to the
series voltage drop, which occurs across the series impedance of the transformer.
Moreover, figures 5.1 to 5.5 show that the voltage profiles for the feeders CC1 to CC5 are
not balanced, more so towards the end of the circuits, with CC4 having the largest series
volt drop. As discussed in chapter 3 of this report, these network voltage imbalances can be
attributed to the presence of many single phase loads, causing different voltage drops
across the carrier phases. A summary of the feeder and district transformer currents is
provided in table 5.1.
62
Table 5.1 Base case district transformer and feeder parameter summaries
Looking at table 5.1, it can be seen that the respective CC1 to CC5 phase currents are not
balanced. This is due to the inability to distribute the single phase loads perfectly across the
three phases of the respective feeders. These imbalances have resulted in zero sequence
currents being present at the connection point of all the feeders, as displayed in table 5.2.
Table 5.2 Base case feeder zero sequence and neutral current summaries
Feeder Zero sequence current (A) aveIIo
,φ Neutral current (A)
CC1 6.61 8% 19.83 CC2 5.07 6% 15.21 CC3 0.02 0% 0.06 CC4 6.57 4% 19.71 CC5 5.4 3% 16.2
In assessing table 5.2 and figures 5.1 to 5.5, it can be evaluated that where the feeder zero
sequence current is large, more so with respect to the average phase current, the voltage
63
imbalances are most severe. This is evident in CC1, where the voltage profile is most
volatile as displayed in figure 5.1. However, looking at the phase currents drawn by CC3, it
is clear that they are fairly well matched in terms of magnitude. This balance has resulted in
a very small zero sequence current, as shown in table 5.2, and the fairly balanced voltage
profile displayed in figure 5.3.
Further to the above discussions, in assessing the results for the busbar and terminal
summaries of appendix C, these phase imbalances become more evident at various points
of the network. This result can be clearly assessed by looking at the power flow summary
diagrams. There, it can be observed that one single phase service connection results in a
higher zero sequence current. The zero sequence current is reduced when the feeder is
attempted to be balanced at the upstream junctions by distributing other single phase
services evenly across the other phases. This effect can be explained by the fact that the
zero sequence current at any point in the network is directly related to the phase currents,
according to the sequence components matrix. Therefore, where there exists a single phase
tee off, the line currents upstream are imbalanced resulting in a higher zero sequence
current. This zero sequence current remains unchanged until this imbalance is reduced by
connecting other single phase loads of similar magnitude to the remaining phases.
Based on this analysis, it is important to avoid large distances between single phase load
connections in a three phase network. If large distances are involved between single phase
load connections, then the voltage drop across the neutral conductor increases, thereby
64
enhancing network voltage imbalances. This may also produce a large enough potential
difference between earth and the neutral that could result in nuisance shocks in households.
5.2.1.2 Grid losses
The grid losses for the base case power flow study have been summarized in table 5.3.
Table 5.3 Base case grid losses Grid losses Real power (kW) 5.64 Reactive power (kVAr) 22.43
The losses in table 5.3 can be attributed to the series impedance of the transformer and the
cables used in the network.
5.2.2 Base case LV network with PV penetration
The effects of having a 1.5kW PV system at every customer service point, on the entire
specified LV network have been investigated. This power flow study has been conducted
with the network load set to its maximum value. The PV systems are scaled to their
ultimate maximum output of 80% relative to the nominal rating, as per the analysis made in
section 4.4.1 of this report.
For this scenario, the integrated PV system capacity relative to the 500kVA district
transformer and the maximum load has been calculated to be 20% and 23% respectively. It
should be noted that this power flow study represents an optimum network setup, in terms
of load distribution, transformer and feeder loadings, carrier parameters, and existing
65
network voltage profile, as well as PV distribution at every single phase load connection
point. This scenario may well represent a future distribution network operation point in the
SWIS.
For full results of the power flow study carried out for this network scenario, please see
appendix C.
5.2.2.1 Voltage profile results and discussions
The voltage profiles for the feeders in the network, namely CC1 to CC5 are represented in
figures 5.6-5.10.
Figure 5.6 Voltage profile CC1 with PV penetration
66
Figure 5.7 Voltage profile CC2 with PV penetration
Figure 5.8 Voltage profile CC3 with PV penetration
67
Figure 5.9 Voltage profile CC4 with PV penetration
Figure 5.10 Voltage profile CC5 with PV penetration
68
In assessing figures 5.6 to 5.10, it is apparent that the network voltages are within the
specified operation limits. Even though the feeder voltages are not balanced, the figures
illustrate improvements in the voltage profiles of CC1 – CC5 in comparison to those
analyzed in section 5.2.1 of this report. Furthermore, the series volt drop across the feeder
carriers is lower in comparison with the base case results. These observations are most
evident in the CC4 voltage profile. The phase voltages in CC4, towards the end of the
feeder are fairly well balanced with higher magnitudes in comparison to the respective base
case voltage profile.
A summary of the feeder and district transformer phase currents is provided in tables 5.4
and 5.5.
Table 5.4 CC4 phase loadings with PV
69
Table 5.5 Neutral current variations with and without PV
Feeder Zero sequence current (A) aveI
Io,φ Neutral
current (A) Variations from base case
CC1 5.05 8% 15.15 -24% CC2 5.01 7% 15.03 -1.2% CC3 0.02 0% 0.06 0.00% CC4 2.19 1% 6.57 -67% CC5 4 3% 12 -26%
Looking at table 5.4 it can be seen that the feeder and transformer phase currents are lower
in comparison to the base case power flow results presented in table 5.1. This can be
attributed to the network off-loading effect of the installed PV systems. Table 5.5 shows
that the zero sequence currents for CC1 to CC5, at the LV frame are also lower in
comparison with the base case results. The improved voltage profile of this network is a
result of the network off-loading effect, hence, the reduced phase and neutral currents for
all the feeders.
5.2.2.2 Grid losses
As displayed in table 5.6, the grid losses for this network scenario are substantially lower in
comparison with the base case power flow results. This is a direct result of the reduced line
currents and hence, the reduction in network resistive and inductive losses.
Table 5.6 Grid loss variances, with and without PV
Grid losses Comparison with base case Real power (kW) 3.62 -36% Reactive power (kVAr) 16.52 -27%
70
Based on this analysis, it can be concluded that under the current rebate scheme program,
promoting the installation of 1.5kW grid connected PV systems, on well balanced networks
with good quality carriers, are not expected to experience any issues. This is true, where the
introduced PV systems are well distributed and, thus, do not result in significant network
imbalances. Indeed, the house supply to which the single phase PV systems are connected
shall be single phase for this conclusion to be valid.
5.3 CC4 sensitivity studies
The sensitivity studies conducted in this project aim to monitor whether variations to
network parameters at specific PV penetration levels cause voltage operation limits to be
breached. The network parameter variations for each study case are well defined. Where
the stated network parameter variation causes the voltage operation limits to be in breach of
the specified limits, the subject PV penetration level is deemed to be sensitive to that
change, and vice versa.
Sensitivity studies have been carried out on CC4. As per the discussions made in section
4.2.1 of this report, this feeder was selected for this purpose as,
It is well loaded, with many single phase customer connections
It extends over 300m away from the district transformer
It has the largest base case series volt drop compared to the other feeders in the
specified network.
71
Based on the above, this feeder can be representative of other networks (including smaller
OH networks).
The sensitivity studies carried out on CC4 aim to monitor changes to voltage profile, with
variations made to:
PV system output and network load due to seasonal effects
The feeder carrier type, load and PV penetration
PV penetration and network balance variances
The results and analysis of these variations are presented in the following sections of this
thesis. The DIgSILENT Power Factory network model for CC4 is in appendix A of this
report.
Table 5.7 summarizes the installed loads and PV systems on each phase, throughout CC4.
The installed PV system capacity ratio with respect to the maximum load, defined as PV
penetration level, is documented in table 5.7.
Table 5.7 CC4 installed load and PV system summary
Phase
'a' Phase
'b' Phase
'c' Sub total
Total load (kVA) 47.87 43.17 43.17 134.2 Number of single phase PV systems 8 7 7 22 1.5kW PV system installed capacity (kW) 12 10.5 10.5 33 PV penetration levels with respect to load 25% 24% 24% 25% 3kW PV system installed capacity (kW) 24 21 21 66 PV penetration levels with respect to load 50% 49% 49% 49%
72
5.3.1 Seasonal PV output and network load variations
As discussed in sections 3.3.1 and 4.4.1 of this report, the network load and PV system
performance are subject to daily and seasonal variations. In this part of the chapter, these
effects, on the LV feeder CC4 of the specified network model, are investigated. This
analysis aims to monitor the changes to the base case CC4 feeder voltage operation limits,
with the seasonal variations in 1.5kW PV systems at every single phase service point. The
worst case scenario, in terms of network voltage profile, is when maximum PV penetration
coincides with low network loads. The power flow studies have therefore been carried out
for the 1pm scaling values (displayed in tables 4.2 and 4.5 of this report) of each month
with and without PV penetration. The studies have been conducted for this time of the day,
as the load scaling factors usually resulted in low feeder loadings and the PV scaling
factors resulted in the best PV system performances, for the selected months of:
July 2008
August 2008
November 2008
January 2009
For full results of the power flow studies carried out in this section of the report, please see
appendix C.
73
5.3.1.1 Voltage profile results and discussions
The CC4 voltage profiles for the case of the network loads and PV systems scaled to 1pm
of the respective months are displayed in figures 5.11-5.14. Results for the 1pm voltage
profiles of the selected months with no PV penetration can be assessed in appendix C.
Figure 5.11 CC4 voltage profile –July 2008, 1pm load and PV scales
74
Figure 5.12 CC4 voltage profile – August 2008, 1pm load and PV scales
Figure 5.13 CC4 voltage profile – November 2008, 1pm load and PV scales
75
Figure 5.14 CC4 voltage profile – January 2009, 1pm load and PV scales
In assessing figures 5.11 to 5.14, it is can be seen that the phase voltage magnitudes are
well matched due to the network off loading effect of the PV systems. Furthermore, the
voltage profiles are very close to the upper operation limits of 1.06pu at all points of the
feeder, CC4. It can therefore be concluded that, because the no load district transformer
secondary voltage is set to 1.06pu, the upper voltage operation limits are likely to be
breached at times throughout the year.
In the previous section, it was established that the base case CC4 voltage profile is well
within the required lower voltage operation limits (see figure 5.4). It is therefore possible
to tap down the district transformer, in order to ensure the voltage profile of CC4 is within
the upper voltage operation limits throughout the year.
76
Based on the above analysis it can be concluded that PV penetration levels are not sensitive
to the variations in the output of the PV systems and loads where:
The existing LV distribution network is well balanced;
The network consists of 240SQmm cable;
The PV systems have been distributed evenly across all phases;
The existing full load network voltage profile allows for district transformer tap
setting reduction (where no load secondary voltage is 1.06pu).
5.3.1.2 Network loadings and grid losses
The power flow results for CC4 loadings, with and without PV penetration are displayed in
figure 5.15.
CC4 power input, with and witout PV penetration
0
10
20
30
40
50
60
Jul-08 Aug-08 Sep-08 Oct-08 Nov-08 Dec-08 Jan-09
Month
Uni
t (x1
)
Feeder Infeed NO PV (kW) Feeder Infeed NO PV (kVAr)Feeer infeed NO PV (kVA)Feeder Infeed PV (kW) Feeder Infeed PV (kVAr)Feeder infeed PV (kVA)
Figure 5.15 CC4 seasonal real and reactive power variations, with and without PV
Figure 5.15 displays substantial off-loading of CC4 in the months of August, November
and January. This network off-loading effect can be related to the voltage profiles of the
77
different months. Where the network off-loading effect displayed in figure 5.15 is
substantial, the voltage profile of the respective month sits closer to the district
transformer’s secondary voltage. It should be noted that the off-loading effect is only
occurring for the feeder’s real power infeed, as the PV systems are operating at unity power
factor. Reactive power requirements of the feeder loads are always met by the grid, thus,
the near unchanged reactive power infeed in to the feeder as seen in figure 5.15. This
results in a ‘poor’ power factor throughout the LV network and at the district transformer.
The variations in grid losses for the different months are displayed in figure 5.16.
CC4 Grid losses, with and witout PV penetration
0
0.5
1
1.5
2
2.5
3
3.5
4
4.5
Jul-08 Aug-08 Sep-08 Oct-08 Nov-08 Dec-08 Jan-09
Month
Unit
(x1)
Grid losses NO PV (kW) Grid losses NO PV (kVAr)Grid losses NO PV (kVA)Grid losses PV (kW) Grid losses PV (kVAr)Grid losses PV (kVA)
Figure 5.16 CC4 seasonal variations in grid losses, with and without PV In figure 5.16, it is clear that the real grid losses are not significant. Nevertheless, the losses
are reduced due to the penetration of the PV systems.
78
5.3.2 Feeder carrier type and network operation limits
In section 5.3.1 of this report, the combination of the variations in CC4 loads and PV
system output was investigated. The analysis was conducted with the network carriers
unchanged from the base case 240SQmm XLPE cables. This section of the report seeks to
investigate the CC4 voltage profile when other carriers that are available for installation in
the WP distribution network are used. The analysis has been carried out using the following
procedure:
1. Change the main (backbone) carrier type
2. Monitor the feeder voltage profile with all loads scaled to 1
3. Monitor the feeder voltage profile with the loads scaled to reflect minimum loading
which coincides with the max PV performance (worst case). Thus, the network
loads were scaled to November 1pm and the PV systems were set to their ultimate
maximum performance of 80% of nominal rating.
The above procedure would give insight into the possible variances in feeder voltage
profile and whether the variances can be managed as per the analysis in section 5.3.1 of this
report.
The investigations have been performed with the CC4 backbone carriers changed to:
7/4.75 Aluminum Alloy Conductor (AAC)
7/3.75 AAC
120SQmm XLPE LV cable
79
The 7/4.75AAC and 7/3.75AAC conductors were selected as they are common in OH
networks. Conductors with smaller diameters in comparison to the ones listed are used in
smaller aerial networks, and substituting them in CC4 would result in their overloading. In
any case, the smaller network voltage profiles that make use of smaller conductors would
be similar to the worst performer of the above carriers. The 120SQmm LV cable was
selected for this analysis as it is the weakest mains cable used in the WP underground
networks. It should be noted that the feeder load balance and PV distribution are still
representing an optimal network setup.
For full results of the power flow studies carried out in this section of the report, please see
appendix C.
5.3.2.1 7/4.75AAC Conductor voltage profile results and discussions
The voltage profiles for the network load scaled to 1 and PV system scaled to zero is
displayed in figure 5.17. The voltage profile of the network load scaled to November 1pm
with the PV systems scaled to 80% of their nominal rating are displayed in figure 5.18.
80
Figure 5.17 CC4 voltage profile, 7/4.75 conductor full load and no PV penetration
Figure 5.18 CC4 voltage profile, 7/4.75AAC low load and max PV output
81
The voltage profiles displayed in figures 5.17 and 5.18 show that the voltages at all points
of the feeder are within the specified limits. In assessing figure 5.18, it is apparent that the
feeder voltage profile is sitting close to the upper voltage operation limits due to the off
loading effect of the PV systems. In addition, the feeder voltage profile with full load,
displayed in figure 5.17, is within the specified lower voltage operation limits. It is
therefore possible to tap the district transformer down and reduce the secondary voltage by
2.5%. This would ensure the CC4 voltage profile remains within the upper voltage
operation limits throughout the year.
Based on the above analysis, it can be concluded that LV network voltage profiles will not
be sensitive to the variations in PV system outputs where,
PV penetration levels are between 20%-25% with respect to network load;
The network load and PV systems are well distributed across the three phases;
7/4.75AAC or bigger conductors form the backbone carriers of the LV feeders;
The existing network voltage profiles allow for district transformer tap setting
reduction (where no load secondary voltage is 1.06pu).
5.3.2.2 7/3.75AAC Conductor voltage profile results and discussions
The voltage profile for the network load scaled to 1 and PV system scaled to zero is
displayed in figure 5.19. The voltage profile of the network load scaled to November 1pm
with the PV systems scaled to 80% of their nominal rating is displayed in figure 5.20.
82
Figure 5.19 CC4 voltage profile, 7/3.75AAC full load and no PV penetration
Figure 5.20 CC4 voltage profile, 7/3.75AAC low load and max PV output
83
In assessing the voltage profiles displayed in figures 5.19 and 5.20, it can be seen that the
feeder voltages are just within the specified network voltage limits. The voltage profile
displayed in figure 5.19, representing the full load conditions, is close to the lower voltage
operation limits. Additionally, the voltage profile presented in figure 5.20 is close to the
upper network operation limits. This highlights the issues associated with the weaker LV
distribution network voltage profiles and the off-loading effect of PV systems. For this
case, it is not possible to tap the district transformer down to ensure the off-loaded feeder
voltage operation limits are not violated. Tapping the transformer down would result in the
full load network voltage profile to be in breach of the lower voltage operation limits.
Based on the above analysis, it can be concluded that where 7/3.75AAC or smaller
conductors form the backbone carriers of a LV network, the feeder voltage profile is
sensitive to the variations in PV system outputs. This implies that PV penetration levels of
20%-25% are not feasible, where the weak OH LV network is well loaded.
5.3.2.3 120SQmm XLPE LV cable voltage profile results and discussions
The voltage profile for the network load scaled to 1 and PV system scaled to zero is
displayed in figure 5.21. The voltage profile of the network load scaled to November 1pm
with the PV systems scaled to 80% of their nominal rating is displayed in figure 5.22.
84
Figure 5.21 CC4 voltage profile, 120SQmm cable full load and no PV penetration
Figure 5.22 CC4 voltage profile, 120SQmm cable, low load and max PV output
85
Figure 5.21 shows that the feeder voltages are within the specified lower voltage operation
limits. On the other hand, figure 5.22 illustrates that the upper network operation limits are
breached at some points of the feeder. In this scenario, it is possible to tap the district
transformer down one or two settings in order to ensure the upper voltage operation limits
are not breached at times of low network loadings and high PV penetration levels.
Based on the above analysis, it can be concluded that the feeder voltage profile will not be
sensitive to the variations in PV system outputs where,
PV penetration levels are between 20%-25% with respect to network load;
The network load and PV systems are well distributed across the three phases;
120SQmm or bigger cables form the backbone carriers of the LV feeders;
The existing network voltage profiles allow for district transformer tap setting
reduction (where no load secondary voltage is 1.06pu).
5.3.2.4 Grid losses comparison for the different carriers
The full load grid losses in CC4, for the various carriers investigated are summarized in
figure 5.23. It can be seen that the stronger carriers have the lowest losses and higher
reactive to real power loss ratios. For example, the 240SQmm LV cable clearly has the
lowest losses and a relatively high reactive to resistive loss ratio. This is the reason behind
the 240SQmm LV cable producing a superior voltage profile in comparison with all other
carriers used in the WP distribution network.
86
Full load grid losses for the various carriers
0
2
4
6
8
10
12
14
16
240SQmm 7/4.75AAC 120SQmm 7/3.75AAC
Carrier type
Units
(x1) Grid losses Full load No PV (kW)
Grid losses Full load No PV (kVAr) Grid losses Full load No PV (kVA)
Figure 5.23 Full load CC4 grid losses for the various carriers
5.3.3 PV penetration and network balance variances
5.3.3.1 Three kW PV system investigations
So far in this project, all the analysis has been carried out with the PV systems having a
nominal rating of 1.5kW. This section seeks to investigate the effects of doubling the PV
system size at every single phase customer connection point to 3kW. Power flow studies
have been conducted with the CC4 main carriers set to 240SQmm and 120SQmm LV
cables, with the network load and PV system output scaled to November 1PM and 0.8
respectively. Please refer to appendix C for detailed power flow results.
Voltage Profile results and discussions
The voltage profiles of the 240SQmm and 120SQmm LV cables are displayed in figures
5.24 and 5.25 respectively.
87
Figure 5.24 CC4 Voltage profile, 3kW PV systems with 240SQmm cable
Figure 5.25 CC4 Voltage profile, 3kW PV systems with 120SQmm cable
88
In assessing figures 5.24 and 5.25, it can be seen that the upper voltage operation limits are
in breach. Moreover, it is clear that the transformer secondary voltage is set to 1.06pu, due
to the transformer primary input voltage of 1pu and tap setting of 3. Doubling the PV
system penetration levels to 49.18 with respect to total load, has resulted in reverse power
flow throughout the local network as well as the district transformer. This is the reason for
the district transformer secondary having the lowest feeder voltage set point and the end of
the feeders having the highest voltage set points, as seen in figures 5.24 and 5.25. This
effect has been highlighted further in figure 5.26, where the CC4 loads were set to zero
while the same PV penetration levels were maintained.
Figure 5.26 CC4 Voltage profile, 3kW PV systems, no load with 120SQmm cable
89
In section 5.2.1 of this report, it was established that the base case CC4 voltage profile
consisting of 240SQmm LV cable is well within the specified voltage operation limits.
Thus, it is possible to tap the transformer down by two settings and hence, accommodate
PV penetration levels of 40%-49%. It can therefore be concluded that the feeder voltage
profile will not be sensitive to the variations in PV system outputs where:
PV penetration levels are between 25%-49% with respect to network load;
The network load and PV systems are well distributed across the three phases;
240SQmm or bigger cables form the backbone carriers of the LV feeders;
The existing network voltage profiles allow for at least two tap setting reductions at
the district transformer (where no load secondary voltage is 1.06pu).
Additionally, it can be concluded that the 120SQmm LV cable voltage profile will not be
within the required limits at penetration levels of 40-49%. Thus, the 120SQmm feeder
cable is sensitive to PV penetration levels higher than 20 to 25% and specific studies need
to be conducted when these are in breach. From this analysis it can also be confirmed that
where 7/4.75AAC or smaller conductors are present in a network, these high PV
penetration levels are not feasible and the conclusions made in section 5.3.2 of this report
should be followed.
5.3.3.2 PV systems on phase ‘a’ scaled to zero
Thus far in this project, the network setups have been well balanced in terms of load and
PV distribution across the three feeder phases. In this section of the report, variations to PV
penetration balance across the three phases of CC4 are investigated. In the endeavor of
90
achieving this, all the 1.5kW PV systems connected to phase ‘a-neutral’ of the feeder have
been put out of service. The balance of the loads across the three phases was kept the same
as per base case LV network setup (displayed in table 5.7). These power flow studies have
been carried out with the network load scaled to 1 and PV system output scaled to 0.8, for
the cases of having 120SQmm and 7/4.75AAC as the backbone feeder carriers.
For full results of the power flow studies carried out in this section of the report, please see
appendix C.
Voltage profile results and analysis
The voltage profiles of the 120SQmm cable and 7/4.75AAC LV conductors are displayed
in figures 5.27 and 5.28 respectively.
91
Figure 5.27 CC4 Voltage profile, zero PV penetration on phase ‘a’, 120SQmm cable
Figure 5.28 CC4 Voltage profile, zero PV penetration on phase ‘a’, 7/4.75AAC
92
Looking at figures 5.27 and 5.28, it is clear that the CC4 voltage profile is unbalanced. This
result can be attributed to the different loadings of the phases, with phase ‘a’ having the
highest load due to the zero PV system output. This can be assessed in table 5.8.
Table 5.8 Phase currents of CC4 at district transformer
Feeder carrier type
Phase 'a' current (A)
Phase 'b' current (A)
Phase 'c' current (A)
120SQmm 206.06 144.24 148.09 7/4.75AAC 205.6 144.34 152.7
Looking at table 5.8 it can be seen that phase loading imbalances are present. Furthermore,
it is clear that the 120SQmm LV cable and the 7/4.75AAC conductor are producing slightly
different phase loading results. This can be attributed to the fact that the cable models have
shunt capacitances whereas the overhead line models have negligible shunt capacitance due
to the low line voltage values. Indeed, the phase imbalances produce large zero sequence
currents, causing, neutral currents to flow throughout the network. This effect is illustrated
in table 5.9.
Table 5.9 Zero sequence currents of CC4 at district transformer
Feeder carrier type
Zero sequence current (A)
IN / I phase ave
Neutral current (A)
120SQmm 18.29 33% 54.87 7/4.75AAC 17.92 32% 53.76
These neutral currents will be present throughout the network. These larger than usual zero
sequence, and neutral currents can cause large voltage drops in the representative zero
sequence impedance network. Large zero sequence voltages in a network cause severe
93
voltage imbalance. This effect is further investigated in the next section of this chapter,
where the combined findings are summarized.
5.3.3.3 PV systems and load on phase b scaled to 1 and 0
As an extension to section 3.3.3.2 of this report, this section seeks to investigate the
consequences of maintaining the introduced CC4 1.5kW single phase PV systems on phase
‘b’ of the feeder only, while setting the load on this phase to zero. The load and PV system
output on the other two phases is scaled to one and zero respectively. The power flow
studies have been conducted for the network backbone feeders set to 120SQmm cable and
7/4.75AAC conductor. These investigations aim to further highlight the issues associated
with distributing PV systems unevenly across a three phase system.
For full results of the power flow studies carried out in this section of the report, please see
appendix C.
Voltage profile results and analysis
The voltage profiles of the 120SQmm cable and 7/4.75AAC LV conductors are displayed
in figures 5.29 and 5.30 respectively.
94
Figure 5.29 CC4 Voltage profile, full PV penetration on phase ‘b’, 120SQmm cable
Figure 5.30 CC4 Voltage profile, full PV penetration on phase ‘b’, 7/4.75AAC
95
Looking at figures 5.29 and 5.30, it is clear that the voltage profiles are extremely
unbalanced. This is due to the intentional load and PV distribution imbalances created
across the feeder. In both voltage profiles, the phase ‘b’ voltage magnitude is well above
the allowable operation limits. The lowest voltage magnitude in phase ‘b’ is l.06pu and is
present at the secondary of the district transformer, as displayed in figures 5.29 and 5.30.
The light load at the connection point and reverse power flow in some parts of phase ‘b’,
has resulted in its lowest voltage of 1.06pu, and the over voltages at other points.
The heavy loading of phases ‘a’ and ‘c’ and the lightly loaded phase ‘b’ are displayed in
table 5.10.
Table 5.10 Phase currents of CC4 at district transformer
Feeder carrier type
Phase 'a' current (A)
Phase 'b' current (A)
Phase 'c' current (A)
120SQmm 219.44 16.63 172.26 7/4.75AAC 204.78 17.04 186.77
The heavy loading of phases ‘a’ and ‘c’ has resulted in the large series volt drops in these
phases. The significant phase loading imbalances throughout the feeder has resulted in a
large zero sequence current to be present in CC4. This can be assessed in table 5.11.
Table 5.11 Zero sequence currents of CC4 at district transformer
Feeder carrier type
Zero sequence current (A)
IN / I phase ave
Neutral current (A)
120SQmm 55.04 124% 165.12 7/4.75AAC 54.33 122% 162.99
96
The resulting neutral current in this scenario is larger than the magnitude of the average
phase currents, as seen in table 5.11. This neutral current is present between the Low
Voltage Frame and C4-J3 of figure 4.6.
Indeed, the zero sequence current produces a zero sequence voltage drop in the
representative zero sequence impedance network. The zero sequence voltage drop also
means a larger neutral voltage, and with it, the possibility of nuisance shocks in house
holds. Furthermore, the zero sequence voltage drop results in higher voltages on the lightly
loaded phase(s) and the opposite effect in the heavily loaded phase(s). These effects can
certainly be observed in figures 5.27 to 5.30.
The analysis carried out in this section and section 5.3.3.2 of this report demonstrates that
PV penetration levels are very sensitive to network balance. Unbalanced networks would
not accommodate the PV penetration levels stated so far in this project. It is unclear what
levels of penetration will be feasible in unbalanced networks, as it is hard to gauge the level
of network balance in the first place.
These findings highlight the importance of distributing loads and hence PV systems evenly
across all phases. Failing to do so will result in the creation of unbalanced networks and
hence, large neutral current flows. The consequential undesirable voltage profile will
presents challenges in integrating PV systems, and other forms of DG.
97
The power flow simulation results analyzed in this section, demonstrate the possibility of a
voltage rise issue associated with 3-phase customer services that have single phase PV
installations. The voltage rise may occur at instances where the PV system is outputting
maximum power to the grid, and the other two phases are drawing power from the grid.
This would result in a large neutral current to be present in the service carrier, between the
grid connection point and the customer main switch board. The magnitude of this neutral
current will depend on the magnitude and direction of the phase currents in the service
wire. Where the neutral, and thus zero sequence current is high, then depending on the
length and thermal ratings of the service carrier, the zero sequence voltage may also get
quiet high. This in turn would result in a voltage rise on the inverter connection phase. The
voltage rise may or may not be sufficient to cause nuisance tripping of the inverter.
Whether the voltage rise does or does not trip the inverter, will depend on the existing
voltage profile at the connection point to the grid, and the magnitude of the voltage rise in
the service carrier to the customer main switch board.
The current carrying capability of the service carriers is also important. This is due to the
fact that under worst case phase loading and PV output scenarios, the neutral current can
get quiet high. Indeed, the service carrier has to handle this high neutral current, or else
serious safety issues will be present at the premises.
98
5.4 Summary of results
A concise summary of the results presented so far in this chapter are presented in this
section. In brief, PV penetration levels of 20%-25% with respect to network load can be
feasible in specific LV networks. The LV networks must consist of UG cables of
120SQmm or better or 7/4.75AAC conductors or better and exhibit the following features:
The existing LV distribution network shall be well balanced;
The network load and PV systems shall be distributed evenly across all phases;
throughout the network;
The existing network voltage profiles shall allow for at least one reduction in
district transformer tap setting.
PV penetration levels of 25% - 49% with respect to network load can be sustainable only
where 240SQmm LV cable is used in networks. This is valid provided the LV networks
exhibit the features mentioned earlier, with the following variation:
The existing network voltage profile shall allow for at least two reductions in the
district transformer tap setting.
The validity of the above summaries is greatly dependant on the overall LV network
balance. Where the phase imbalances are large, as discussed in sections 5.3.3.2 and 5.3.3.3,
the specified PV penetration levels would no longer be feasible. Phase loading imbalances
cause larger neutral currents, zero sequence voltages and undesirable voltage profiles.
99
The results presented in section 5.3.3.3 also highlight the possible issues associated with
three phase customer services that have single phase PV installations. These issues are
outlined below:
Possibility of higher than usual neutral currents, between the grid connection point
and the customer main switch board, depending on the PV system performance and
the phase loadings of the service wire;
Possibility of a voltage rise on the inverter connection phase, depending on the
properties and length of the service carrier;
The voltage rise may trip the inverter, depending on its magnitude and the existing
voltage profile at the connection point;
The service carrier has to be rated for the largest possible neutral current, otherwise
serious safety issues will be present at the premises.
100
6 Conclusion
6.1 Conclusions
An investigation of WP’s LV network voltage operation limits with the integration of PV
systems has been completed in this project. The investigation was performed on a typical
WP LV network model, in DIgSILENT Power Factory using WP residential network
loadings data along with solar radiation and temperature data. The analysis of the results,
focused on the issues associated with the WP LV network voltages while highlighting the
variations in grid losses and network loadings. The main findings are summarized:
1. PV penetration levels of 20%-25% can be feasible in LV UG networks consisting of
120SQmm cables or better or OH networks consisting of 7/4.75AAC conductors or
better.
2. PV penetration levels of 25% - 49% can only be sustained in UG LV networks
consisting of 240SQmm LV cable.
3. The validity of the specified feasible PV penetration levels is heavily dependant on
LV network balance, existing network voltage profile and district transformer tap
setting.
4. Although not significant, grid losses are reduced due to the real power offloading
effect of the grid connected PV systems.
5. Poor LV network power factors can result as inverters operate at unity power factor
and do not provide the reactive power requirements of the network loads.
101
6. High neutral currents in service wires and the nuisance tripping of inverters are the
issues associated with a single phase PV system installation in a three phase supply
arrangement.
From findings of this project, it can be concluded that all LV networks in the SWIS have a
limit with respect to PV penetration levels, before the network voltages get adversely
affected. They highlight that the WP UG networks are best suited in adopting grid
connected PV systems. Conversely, the well loaded weak OH network voltage profiles
present challenges in integrating PV system penetration levels of 20-25%. These include
networks with unbalanced load distribution across the three phases that consist of poor
conductors with long distances.
This project also demonstrates that PV systems are not offsetting the peaky residential
network load profile. Therefore, the relevant WP distribution and transmission upgrade
projects can not be deferred as the residential SWIS peak load is effectively unchanged.
The intermittency of the solar resource means that, high grid connected PV system
penetration, with no electrical storage and demand management, provides substantially
lower grid support value when compared to dispatchable DG technologies.
102
6.2 Recommendations and future work
6.2.1 Single phase PV system connection for three phase customers
As discussed previously, the issues associated with single phase PV system installations in
three phase supply arrangements are high neutral currents and nuisance tripping of the
inverter. According to AS4777 and Western Power’s Technical Rules, single phase PV
systems up to 10kW can be connected to the LV distribution grid. Clearly, this limit will
enhance the aforementioned issues. Furthermore, in small LV networks, one 10kVA PV
system will result in high penetration levels, possibly high neutral currents and voltage
imbalance. It is therefore recommended for WP to simulate, test and re-define realistic PV
system size limitations for single phase and three phase customer supply arrangements.
Where the customers request to connect PV systems larger than those specified by WP,
specific network analysis shall be made in approving the connection. The network analysis
may suggest to reinforce a network prior to the connection or to simply connect the inverter
to the customer phase that is well loaded throughout the day.
6.2.2 LV network balance in new subdivisions
When new subdivisions are electrically designed, it is unknown whether the customer
supply arrangement is single phase or three phase. Moreover, the task of distributing the
single phase customer supplies evenly across all phases can become challenging as not all
the landowners build and move in simultaneously. It is worthwhile to review the current
practice in WP to ensure the single phase loads in new subdivisions is well distributed.
Another possibility is for WP to lower the limits of power that can be supplied in a single
103
phase supply arrangement. It is also important to ensure electricians are distributing
household loads evenly across a three phase supply arrangement.
6.2.3 The effect of offloading or reverse power flow in district transformers
The load flow studies in this project showed that poor network power factors can occur as a
result of the inverters operating at unity power factor. Are there any issues associated with
this effect, keeping in mind that the apparent power supplied by the transformer and the
grid losses would be lower than otherwise? In addition, are the district transformers which
are used in the WP distribution network capable of delivering power from LV to HV
networks (reverse power flows)?
6.2.4 Demand management
The findings of this project highlighted that the peak residential network load does not
coincide with the best PV system performance, throughout a year. Based on this, to what
degree will demand management in conjunction with grid connected PV systems reduce the
stress on some of the SWIS LV networks?
104
7 References 1. Government of Western Australia Office of Energy, Strategic Energy Initiative
Issues Paper. 2009: Perth Western Australia. p. 16.
2. Western Power, Submission to the Strategic states Energy Initiative’s and Issues Paper. 2010. p. 25.
3. Puttgen, H.B., P.R. MacGregor, and F.C. Lambert, Distributed generation: Semantic hype or the dawn of a new era? Power and Energy Magazine, IEEE, 2003. 1(1): p. 22-29.
4. Caisheng, W. and M.H. Nehrir, Analytical approaches for optimal placement of distributed generation sources in power systems. Power Systems, IEEE Transactions on, 2004. 19(4): p. 2068-2076.
5. Brown, R.E. and L.A.A. Freeman. Analyzing the reliability impact of distributed generation. in Power Engineering Society Summer Meeting, 2001. IEEE. 2001.
6. Rawson, M., Distributed Generation Costs and Benefits Issue Paper. Public Interest Energy Research California Energy Commission, 2004: p. 1-24.
7. Magueed, F. and J. Daalder. Parallel Operation of Distributed Generation in Weak Distribution Systems. in Power Electronics and Motion Control Conference, 2006. EPE-PEMC 2006. 12th International. 2006.
8. Paska, J. and A. Oleksy. Reliability issues in power systems with DG. in Power and Energy Society General Meeting - Conversion and Delivery of Electrical Energy in the 21st Century, 2008 IEEE. 2008.
9. G. Pepermans, et al., Distributed Generation: Definition, Benefits and Issues. K.U.Leuven - Energy Institute, 2003: p. 1-22.
10. Moran, A., The Electricity Industry in Australia: Problems Along the Way to a National Electricity Market. Institute of Public Affairs, Melbourne, Australia, 2007: p. 173-202.
105
11. Lee, W.J. and C.H. Lin. Utility deregulation and its impact on the industrial power systems. in Industrial and Commercial Power Systems Technical Conference, 1997. Conference Record, Papers Presented at the 1997 Annual Meeting., IEEE 1997. 1997.
12. Le, A.D.T., et al. Optimal Distributed Generation Parameters for Reducing Losses with Economic Consideration. in Power Engineering Society General Meeting, 2007. IEEE. 2007.
13. Watt, M., National Survey Report of PV Power Applications Australia, University of NSW: Sydney, Australia. p. 24.
14. Kirti Mala, August Schla" pfer, and T. Pryor, Better or worse? The role of solar photovoltaic (PV) systems in sustainable development: Case studies of remote atoll communities in Kiribati. Renewable Energy, 2008: p. 358-362.
15. IEA Photovoltaics Power Systems Programme. Trends in PV Applications. 2009 [cited 10/11/2009]; Available from: http://www.iea-pvps.org/.
16. Solar Online Australia. 2001 [cited 10/11/2009]; Available from: http://www.solaronline.com.au/solar_rebate_info/.
17. Western Power, Distributed Facilities Management System (DFIS). 2009, Western Power.
18. Bureau of Meterology. Average daily solar exposure Annual. [cited 13, 12, 2009]; Available from: http://www.bom.gov.au/jsp/ncc/climate_averages/solar-exposure/index.jsp.
19. NASA. NASA Surface meteorology and Solar Energy. [cited 13/12/2009]; Available from: http://eosweb.larc.nasa.gov/cgi-bin/sse/grid.cgi.
20. AlHajri, M.F. and M.E. El-Hawary. The effect of Distributed Generation modeling and static load representation on the optimal integrated sizing and network losses. in Electrical and Computer Engineering, 2008. CCECE 2008. Canadian Conference on. 2008.
106
21. Favuzza, S., et al. Comparison of power quality impact of different photovoltaic inverters: the viewpoint of the grid. in Industrial Technology, 2004. IEEE ICIT '04. 2004 IEEE International Conference on. 2004.
22. Tom Butler and W. Peter, Impacts and Benefits of Highly Distributed Embedded Generation in Australian Electricity Distribution Networks, in Econnect Poject NO:2175, T. Moron, Editor. 2009, Econnect: New South Wales. p. 197.
23. Prakasit Sritakaew, Anawach Sangswang, and K. Kirtikara, On the Reliability Improvement of Distribution Systems Using PV Grid-Connected Systems. ECTI TRANSACTIONS ON ELECTRICAL ENG., ELECTRONICS, AND COMMUNICATIONS, 2007. VOL.5(NO.1): p. 79-85.
24. Hoff, T. and D.S. Shugar, The value of grid-support photovoltaics in reducing distribution system losses. Energy Conversion, IEEE Transactions on, 1995. 10(3): p. 569-576.
25. Lawrence J. Borle, Michael S. Dymond, and C.V. Nayar, Development and Testing of a 20-kW Grid Interactive Photovoltaic Power Conditioning System in Western Australia. IEEE TRANSACTIONS ON INDUSTRY APPLICATIONS, 1997. 33(2): p. 502-508.
26. Thomas S. Basso, High-Penetration, Grid-Connected Photovoltaic Technology Codes and Standards, in 33rd IEEE Photovoltaic Specialists Conference. 2008, NREL National Renewable Energy Laboratory: San Diego, California. p. 7.
27. Kieny, C., et al. Distribution grid security management with high DG penetration rate: Situation in France and some future trends. in Power and Energy Society General Meeting - Conversion and Delivery of Electrical Energy in the 21st Century, 2008 IEEE. 2008.
28. Hermann Laukamp, et al., Grid Effects of a Large, Distributed PV Capacity in a Modern Urban Settelement. Fraunhofer Institute Institute for Solar Energy Systems ISE, 2007: p. 1-4.
29. Ryoichi Hara, et al., Demonstration Grid-Connected Photovoltaic Projects in Japan, in IEEE Power and energy magazine. 2009. p. 77-85.
30. Western Power, Technial Rules for the South West Interconnected Network, W. Power, Editor. 2007: Perth Western Australia. p. 100-106.
107
31. Standard Australia, Grid connection of energy systems via inverters, in Part 3: Grid protection requirements, S. Australia, Editor. 2005, Standard Australia: Australia. p. 7.
32. High Penetration of Photovoltaic (PV) Systems into the Distribution Grid. 2009, U.S Department of Energy: Ontario, CA. p. 182.
33. Western Power, Distribution Substation Manual. 2007, Western Power: Perth, Western Australia. p. 6-8.
34. Energy Authority of New South Wales, Earthing Handbook, ed. Energy Authority of New South Wales. Vol. 1. 1986, Sydney: Energy Authority of New South Wales
35. DIgSILENT GmbH, Power Factory User's Manual. 2008, DIgSILENT GmbH: Gormaringen, Germany. p. 1-1.
36. SMA Solar Technology. Sunny Boy. [cited 19/09/2009]; Available from: http://www.sma-australia.com.au/en_AU/products/solar-inverters/sunny-boy.html.
37. T. VU VAN A, et al., Impacts of Ditributed Generation on Distribution System Power Quality, in Electrical Power Quality and Utilisation. 2003: Cracow, Poland. p. 8.
38. Western Power, Amherst Road Canningvale. 2009, [MS013617 Internal Western Power document]: Perth Western Australia.
39. Western Power, Underground Distribution Schemes. 1995, Western Power: Perth, Western Australia. p. 138.
40. Western Power. Residential ADMD calculator. [cited 01/11/2009]; Available from: http://www.westernpower.com.au/mainContent/connectionsUpgrades/policiesRegulations/residentialcalculator.html.
41. Real Estate.com.au. [cited 01/11/2009]; Available from: http://www.realestate.com.au/buy?preferredState=WA.
108
42. Bureau of Meterology. Mean maximum temperature. [cited 8/02/2010]; Available from: http://www.bom.gov.au/jsp/ncc/cdio/cvg/av?p_stn_num=009225&p_prim_element_index=0&p_display_type=statGraph&period_of_avg=ALL&normals_years=allYearOfData&staticPage=.
43. GE Network Solutions, ENMAC Webview. 2009, Western Power: Perth, WA.
44. Fortis. On-Line Weather Station. 1998 [cited 03, 12, 2009]; Available from: http://wwwmet.murdoch.edu.au/.
45. Honsberg, C. and S. Bowden, Photovoltaics: Devices, Systems and Applications. 1998, The University of NSW Photovoltaics Centre: Sydney.
46. Australian and New Zealand Solar Energy Society, Australian Solar Radiation Handbook. 2006, Australian and New Zealand Solar Energy Society.
109
Appendix A Load flow model Appendix A is located in the folder named “Appendices” in the root directory of the
enclosed compact disc. It consists of the ENG460 Thesis Project DIgSILENT Power
Factory model. In the model there are two study cases, one for the entire LV network
labeled “LV Model development” and another for CC4 studies named “Circuit 4 detail”.
The file can be imported in to Power Factory version 14 and there, the details of the various
WP carrier and transformer models which have been created can be assessed.
110
Appendix B Load and solar radiation data
Appendix B is located in the folder named “Appendices” in the root directory of the
enclosed compact disc. It contains the Microsoft Excel (97-2003) files presenting the Webb
Street HV feeder data, solar radiation and temperature data and the selected LV network
and PV system scaling factors with statistical analysis. It also contains the daily hourly
feeder loadings and solar radiation plots for each month for (2008/2009).
111
Appendix C Power flow results
Appendix C is located in the folder named “Appendices” in the root directory of the
enclosed compact disc. It contains voltage profiles (graphic and descriptive), busbar and
terminal summaries (power flows, power factor, voltage and current), power flow diagram,
and system summary (system losses and total power input). The files are all in .pdf format
and require Adobe Reader (basic version) for access.